Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +4 -0
- added_tokens.json +3 -0
- chat_template.jinja +47 -0
- checkpoint-2000/added_tokens.json +3 -0
- checkpoint-2000/chat_template.jinja +47 -0
- checkpoint-2000/config.json +56 -0
- checkpoint-2000/generation_config.json +11 -0
- checkpoint-2000/model.safetensors +3 -0
- checkpoint-2000/optimizer.pt +3 -0
- checkpoint-2000/rng_state.pth +3 -0
- checkpoint-2000/scheduler.pt +3 -0
- checkpoint-2000/special_tokens_map.json +33 -0
- checkpoint-2000/tokenizer.json +3 -0
- checkpoint-2000/tokenizer.model +3 -0
- checkpoint-2000/tokenizer_config.json +0 -0
- checkpoint-2000/trainer_state.json +474 -0
- checkpoint-2000/training_args.bin +3 -0
- checkpoint-2200/added_tokens.json +3 -0
- checkpoint-2200/chat_template.jinja +47 -0
- checkpoint-2200/config.json +56 -0
- checkpoint-2200/generation_config.json +11 -0
- checkpoint-2200/model.safetensors +3 -0
- checkpoint-2200/optimizer.pt +3 -0
- checkpoint-2200/rng_state.pth +3 -0
- checkpoint-2200/scheduler.pt +3 -0
- checkpoint-2200/special_tokens_map.json +33 -0
- checkpoint-2200/tokenizer.json +3 -0
- checkpoint-2200/tokenizer.model +3 -0
- checkpoint-2200/tokenizer_config.json +0 -0
- checkpoint-2200/trainer_state.json +518 -0
- checkpoint-2200/training_args.bin +3 -0
- checkpoint-2250/added_tokens.json +3 -0
- checkpoint-2250/chat_template.jinja +47 -0
- checkpoint-2250/config.json +56 -0
- checkpoint-2250/generation_config.json +11 -0
- checkpoint-2250/model.safetensors +3 -0
- checkpoint-2250/optimizer.pt +3 -0
- checkpoint-2250/rng_state.pth +3 -0
- checkpoint-2250/scheduler.pt +3 -0
- checkpoint-2250/special_tokens_map.json +33 -0
- checkpoint-2250/tokenizer.json +3 -0
- checkpoint-2250/tokenizer.model +3 -0
- checkpoint-2250/tokenizer_config.json +0 -0
- checkpoint-2250/trainer_state.json +525 -0
- checkpoint-2250/training_args.bin +3 -0
- config.json +56 -0
- generation_config.json +11 -0
- model.safetensors +3 -0
- special_tokens_map.json +33 -0
- tokenizer.json +3 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
checkpoint-2000/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
checkpoint-2200/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
checkpoint-2250/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<image_soft_token>": 262144
|
| 3 |
+
}
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{{ bos_token }}
|
| 2 |
+
{%- if messages[0]['role'] == 'system' -%}
|
| 3 |
+
{%- if messages[0]['content'] is string -%}
|
| 4 |
+
{%- set first_user_prefix = messages[0]['content'] + '
|
| 5 |
+
|
| 6 |
+
' -%}
|
| 7 |
+
{%- else -%}
|
| 8 |
+
{%- set first_user_prefix = messages[0]['content'][0]['text'] + '
|
| 9 |
+
|
| 10 |
+
' -%}
|
| 11 |
+
{%- endif -%}
|
| 12 |
+
{%- set loop_messages = messages[1:] -%}
|
| 13 |
+
{%- else -%}
|
| 14 |
+
{%- set first_user_prefix = "" -%}
|
| 15 |
+
{%- set loop_messages = messages -%}
|
| 16 |
+
{%- endif -%}
|
| 17 |
+
{%- for message in loop_messages -%}
|
| 18 |
+
{%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
|
| 19 |
+
{{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
|
| 20 |
+
{%- endif -%}
|
| 21 |
+
{%- if (message['role'] == 'assistant') -%}
|
| 22 |
+
{%- set role = "model" -%}
|
| 23 |
+
{%- else -%}
|
| 24 |
+
{%- set role = message['role'] -%}
|
| 25 |
+
{%- endif -%}
|
| 26 |
+
{{ '<start_of_turn>' + role + '
|
| 27 |
+
' + (first_user_prefix if loop.first else "") }}
|
| 28 |
+
{%- if message['content'] is string -%}
|
| 29 |
+
{{ message['content'] | trim }}
|
| 30 |
+
{%- elif message['content'] is iterable -%}
|
| 31 |
+
{%- for item in message['content'] -%}
|
| 32 |
+
{%- if item['type'] == 'image' -%}
|
| 33 |
+
{{ '<start_of_image>' }}
|
| 34 |
+
{%- elif item['type'] == 'text' -%}
|
| 35 |
+
{{ item['text'] | trim }}
|
| 36 |
+
{%- endif -%}
|
| 37 |
+
{%- endfor -%}
|
| 38 |
+
{%- else -%}
|
| 39 |
+
{{ raise_exception("Invalid content type") }}
|
| 40 |
+
{%- endif -%}
|
| 41 |
+
{{ '<end_of_turn>
|
| 42 |
+
' }}
|
| 43 |
+
{%- endfor -%}
|
| 44 |
+
{%- if add_generation_prompt -%}
|
| 45 |
+
{{'<start_of_turn>model
|
| 46 |
+
'}}
|
| 47 |
+
{%- endif -%}
|
checkpoint-2000/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<image_soft_token>": 262144
|
| 3 |
+
}
|
checkpoint-2000/chat_template.jinja
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{{ bos_token }}
|
| 2 |
+
{%- if messages[0]['role'] == 'system' -%}
|
| 3 |
+
{%- if messages[0]['content'] is string -%}
|
| 4 |
+
{%- set first_user_prefix = messages[0]['content'] + '
|
| 5 |
+
|
| 6 |
+
' -%}
|
| 7 |
+
{%- else -%}
|
| 8 |
+
{%- set first_user_prefix = messages[0]['content'][0]['text'] + '
|
| 9 |
+
|
| 10 |
+
' -%}
|
| 11 |
+
{%- endif -%}
|
| 12 |
+
{%- set loop_messages = messages[1:] -%}
|
| 13 |
+
{%- else -%}
|
| 14 |
+
{%- set first_user_prefix = "" -%}
|
| 15 |
+
{%- set loop_messages = messages -%}
|
| 16 |
+
{%- endif -%}
|
| 17 |
+
{%- for message in loop_messages -%}
|
| 18 |
+
{%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
|
| 19 |
+
{{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
|
| 20 |
+
{%- endif -%}
|
| 21 |
+
{%- if (message['role'] == 'assistant') -%}
|
| 22 |
+
{%- set role = "model" -%}
|
| 23 |
+
{%- else -%}
|
| 24 |
+
{%- set role = message['role'] -%}
|
| 25 |
+
{%- endif -%}
|
| 26 |
+
{{ '<start_of_turn>' + role + '
|
| 27 |
+
' + (first_user_prefix if loop.first else "") }}
|
| 28 |
+
{%- if message['content'] is string -%}
|
| 29 |
+
{{ message['content'] | trim }}
|
| 30 |
+
{%- elif message['content'] is iterable -%}
|
| 31 |
+
{%- for item in message['content'] -%}
|
| 32 |
+
{%- if item['type'] == 'image' -%}
|
| 33 |
+
{{ '<start_of_image>' }}
|
| 34 |
+
{%- elif item['type'] == 'text' -%}
|
| 35 |
+
{{ item['text'] | trim }}
|
| 36 |
+
{%- endif -%}
|
| 37 |
+
{%- endfor -%}
|
| 38 |
+
{%- else -%}
|
| 39 |
+
{{ raise_exception("Invalid content type") }}
|
| 40 |
+
{%- endif -%}
|
| 41 |
+
{{ '<end_of_turn>
|
| 42 |
+
' }}
|
| 43 |
+
{%- endfor -%}
|
| 44 |
+
{%- if add_generation_prompt -%}
|
| 45 |
+
{{'<start_of_turn>model
|
| 46 |
+
'}}
|
| 47 |
+
{%- endif -%}
|
checkpoint-2000/config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_sliding_window_pattern": 6,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Gemma3ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_logit_softcapping": null,
|
| 9 |
+
"bos_token_id": 2,
|
| 10 |
+
"cache_implementation": "hybrid",
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"final_logit_softcapping": null,
|
| 13 |
+
"head_dim": 256,
|
| 14 |
+
"hidden_activation": "gelu_pytorch_tanh",
|
| 15 |
+
"hidden_size": 640,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 2048,
|
| 18 |
+
"layer_types": [
|
| 19 |
+
"sliding_attention",
|
| 20 |
+
"sliding_attention",
|
| 21 |
+
"sliding_attention",
|
| 22 |
+
"sliding_attention",
|
| 23 |
+
"sliding_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"sliding_attention",
|
| 26 |
+
"sliding_attention",
|
| 27 |
+
"sliding_attention",
|
| 28 |
+
"sliding_attention",
|
| 29 |
+
"sliding_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"sliding_attention",
|
| 32 |
+
"sliding_attention",
|
| 33 |
+
"sliding_attention",
|
| 34 |
+
"sliding_attention",
|
| 35 |
+
"sliding_attention",
|
| 36 |
+
"full_attention"
|
| 37 |
+
],
|
| 38 |
+
"max_position_embeddings": 32768,
|
| 39 |
+
"model_type": "gemma3_text",
|
| 40 |
+
"num_attention_heads": 4,
|
| 41 |
+
"num_hidden_layers": 18,
|
| 42 |
+
"num_key_value_heads": 1,
|
| 43 |
+
"pad_token_id": 0,
|
| 44 |
+
"query_pre_attn_scalar": 256,
|
| 45 |
+
"rms_norm_eps": 1e-06,
|
| 46 |
+
"rope_local_base_freq": 10000.0,
|
| 47 |
+
"rope_scaling": null,
|
| 48 |
+
"rope_theta": 1000000.0,
|
| 49 |
+
"sliding_window": 512,
|
| 50 |
+
"sliding_window_pattern": 6,
|
| 51 |
+
"torch_dtype": "bfloat16",
|
| 52 |
+
"transformers_version": "4.52.4",
|
| 53 |
+
"use_bidirectional_attention": false,
|
| 54 |
+
"use_cache": true,
|
| 55 |
+
"vocab_size": 262144
|
| 56 |
+
}
|
checkpoint-2000/generation_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cache_implementation": "hybrid",
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
1,
|
| 6 |
+
106
|
| 7 |
+
],
|
| 8 |
+
"top_k": 64,
|
| 9 |
+
"top_p": 0.95,
|
| 10 |
+
"transformers_version": "4.52.4"
|
| 11 |
+
}
|
checkpoint-2000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7204dfb6f3a944a032ca5ed20d71c923a4fed2c3b11dfabc2c1ff9f2fabe8af8
|
| 3 |
+
size 536223056
|
checkpoint-2000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ca1a55e9b06ab065a794a5a279c828b57c60635010bf66dec7ffd15ff01ee57
|
| 3 |
+
size 1072590714
|
checkpoint-2000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2ec28ea0c416565eeac14a0e9c944f185ac250f4ed4bd15c84ff77ed78ba9301
|
| 3 |
+
size 14244
|
checkpoint-2000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3acb1a08fa90ba7cf2a95ded566b5ae5aa74d9f58bbee28804cc9682a3227ce
|
| 3 |
+
size 1064
|
checkpoint-2000/special_tokens_map.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"boi_token": "<start_of_image>",
|
| 3 |
+
"bos_token": {
|
| 4 |
+
"content": "<bos>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"eoi_token": "<end_of_image>",
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"content": "<eos>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false
|
| 17 |
+
},
|
| 18 |
+
"image_token": "<image_soft_token>",
|
| 19 |
+
"pad_token": {
|
| 20 |
+
"content": "<pad>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false
|
| 25 |
+
},
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": false,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
}
|
| 33 |
+
}
|
checkpoint-2000/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e570d6288ff0afcea981a80492eddfa3e2239a79de89e5074cbb74b548fa5e2b
|
| 3 |
+
size 33384833
|
checkpoint-2000/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
|
| 3 |
+
size 4689074
|
checkpoint-2000/tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-2000/trainer_state.json
ADDED
|
@@ -0,0 +1,474 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2000,
|
| 3 |
+
"best_metric": 1.8876391649246216,
|
| 4 |
+
"best_model_checkpoint": "./gemma3-270m-turkish_instructions-finetuned/checkpoint-2000",
|
| 5 |
+
"epoch": 1.7777777777777777,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 2000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.044444444444444446,
|
| 14 |
+
"grad_norm": 19.25,
|
| 15 |
+
"learning_rate": 1.088888888888889e-05,
|
| 16 |
+
"loss": 3.5828,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.08888888888888889,
|
| 21 |
+
"grad_norm": 17.875,
|
| 22 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 23 |
+
"loss": 2.4877,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.08888888888888889,
|
| 28 |
+
"eval_loss": 2.275900363922119,
|
| 29 |
+
"eval_runtime": 48.3609,
|
| 30 |
+
"eval_samples_per_second": 20.678,
|
| 31 |
+
"eval_steps_per_second": 10.339,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.13333333333333333,
|
| 36 |
+
"grad_norm": 18.25,
|
| 37 |
+
"learning_rate": 3.311111111111112e-05,
|
| 38 |
+
"loss": 2.255,
|
| 39 |
+
"step": 150
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.17777777777777778,
|
| 43 |
+
"grad_norm": 13.4375,
|
| 44 |
+
"learning_rate": 4.422222222222222e-05,
|
| 45 |
+
"loss": 2.2559,
|
| 46 |
+
"step": 200
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.17777777777777778,
|
| 50 |
+
"eval_loss": 2.1721303462982178,
|
| 51 |
+
"eval_runtime": 48.5933,
|
| 52 |
+
"eval_samples_per_second": 20.579,
|
| 53 |
+
"eval_steps_per_second": 10.289,
|
| 54 |
+
"step": 200
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.2222222222222222,
|
| 58 |
+
"grad_norm": 18.875,
|
| 59 |
+
"learning_rate": 4.99826726554013e-05,
|
| 60 |
+
"loss": 2.2218,
|
| 61 |
+
"step": 250
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.26666666666666666,
|
| 65 |
+
"grad_norm": 16.625,
|
| 66 |
+
"learning_rate": 4.983543173414964e-05,
|
| 67 |
+
"loss": 2.1805,
|
| 68 |
+
"step": 300
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26666666666666666,
|
| 72 |
+
"eval_loss": 2.1423935890197754,
|
| 73 |
+
"eval_runtime": 48.7813,
|
| 74 |
+
"eval_samples_per_second": 20.5,
|
| 75 |
+
"eval_steps_per_second": 10.25,
|
| 76 |
+
"step": 300
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.3111111111111111,
|
| 80 |
+
"grad_norm": 15.0625,
|
| 81 |
+
"learning_rate": 4.953882760420223e-05,
|
| 82 |
+
"loss": 2.2345,
|
| 83 |
+
"step": 350
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.35555555555555557,
|
| 87 |
+
"grad_norm": 15.75,
|
| 88 |
+
"learning_rate": 4.909464407769633e-05,
|
| 89 |
+
"loss": 2.1545,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.35555555555555557,
|
| 94 |
+
"eval_loss": 2.081547260284424,
|
| 95 |
+
"eval_runtime": 48.8759,
|
| 96 |
+
"eval_samples_per_second": 20.46,
|
| 97 |
+
"eval_steps_per_second": 10.23,
|
| 98 |
+
"step": 400
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.4,
|
| 102 |
+
"grad_norm": 17.875,
|
| 103 |
+
"learning_rate": 4.850555252662495e-05,
|
| 104 |
+
"loss": 2.0966,
|
| 105 |
+
"step": 450
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.4444444444444444,
|
| 109 |
+
"grad_norm": 13.1875,
|
| 110 |
+
"learning_rate": 4.7775095816891336e-05,
|
| 111 |
+
"loss": 2.0782,
|
| 112 |
+
"step": 500
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.4444444444444444,
|
| 116 |
+
"eval_loss": 2.0508856773376465,
|
| 117 |
+
"eval_runtime": 48.9079,
|
| 118 |
+
"eval_samples_per_second": 20.447,
|
| 119 |
+
"eval_steps_per_second": 10.223,
|
| 120 |
+
"step": 500
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.4888888888888889,
|
| 124 |
+
"grad_norm": 15.9375,
|
| 125 |
+
"learning_rate": 4.690766700109659e-05,
|
| 126 |
+
"loss": 2.115,
|
| 127 |
+
"step": 550
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.5333333333333333,
|
| 131 |
+
"grad_norm": 13.9375,
|
| 132 |
+
"learning_rate": 4.590848289820442e-05,
|
| 133 |
+
"loss": 2.0314,
|
| 134 |
+
"step": 600
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.5333333333333333,
|
| 138 |
+
"eval_loss": 2.0262327194213867,
|
| 139 |
+
"eval_runtime": 48.768,
|
| 140 |
+
"eval_samples_per_second": 20.505,
|
| 141 |
+
"eval_steps_per_second": 10.253,
|
| 142 |
+
"step": 600
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.5777777777777777,
|
| 146 |
+
"grad_norm": 14.5,
|
| 147 |
+
"learning_rate": 4.4783552718978e-05,
|
| 148 |
+
"loss": 2.0892,
|
| 149 |
+
"step": 650
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.6222222222222222,
|
| 153 |
+
"grad_norm": 14.0,
|
| 154 |
+
"learning_rate": 4.3539641925879495e-05,
|
| 155 |
+
"loss": 2.0352,
|
| 156 |
+
"step": 700
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.6222222222222222,
|
| 160 |
+
"eval_loss": 1.9995155334472656,
|
| 161 |
+
"eval_runtime": 48.8752,
|
| 162 |
+
"eval_samples_per_second": 20.46,
|
| 163 |
+
"eval_steps_per_second": 10.23,
|
| 164 |
+
"step": 700
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.6666666666666666,
|
| 168 |
+
"grad_norm": 13.25,
|
| 169 |
+
"learning_rate": 4.2184231544782596e-05,
|
| 170 |
+
"loss": 2.0117,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.7111111111111111,
|
| 175 |
+
"grad_norm": 13.25,
|
| 176 |
+
"learning_rate": 4.072547317320281e-05,
|
| 177 |
+
"loss": 1.9848,
|
| 178 |
+
"step": 800
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.7111111111111111,
|
| 182 |
+
"eval_loss": 1.9770301580429077,
|
| 183 |
+
"eval_runtime": 48.7445,
|
| 184 |
+
"eval_samples_per_second": 20.515,
|
| 185 |
+
"eval_steps_per_second": 10.258,
|
| 186 |
+
"step": 800
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.7555555555555555,
|
| 190 |
+
"grad_norm": 13.375,
|
| 191 |
+
"learning_rate": 3.9172139955630774e-05,
|
| 192 |
+
"loss": 2.0014,
|
| 193 |
+
"step": 850
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.8,
|
| 197 |
+
"grad_norm": 12.75,
|
| 198 |
+
"learning_rate": 3.7533573820809006e-05,
|
| 199 |
+
"loss": 1.9938,
|
| 200 |
+
"step": 900
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.8,
|
| 204 |
+
"eval_loss": 1.9461404085159302,
|
| 205 |
+
"eval_runtime": 48.8094,
|
| 206 |
+
"eval_samples_per_second": 20.488,
|
| 207 |
+
"eval_steps_per_second": 10.244,
|
| 208 |
+
"step": 900
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.8444444444444444,
|
| 212 |
+
"grad_norm": 13.8125,
|
| 213 |
+
"learning_rate": 3.5819629298273245e-05,
|
| 214 |
+
"loss": 1.9914,
|
| 215 |
+
"step": 950
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.8888888888888888,
|
| 219 |
+
"grad_norm": 13.6875,
|
| 220 |
+
"learning_rate": 3.4040614252052305e-05,
|
| 221 |
+
"loss": 1.9724,
|
| 222 |
+
"step": 1000
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.8888888888888888,
|
| 226 |
+
"eval_loss": 1.9275351762771606,
|
| 227 |
+
"eval_runtime": 48.8023,
|
| 228 |
+
"eval_samples_per_second": 20.491,
|
| 229 |
+
"eval_steps_per_second": 10.245,
|
| 230 |
+
"step": 1000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.9333333333333333,
|
| 234 |
+
"grad_norm": 13.25,
|
| 235 |
+
"learning_rate": 3.2207227887960935e-05,
|
| 236 |
+
"loss": 1.9607,
|
| 237 |
+
"step": 1050
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.9777777777777777,
|
| 241 |
+
"grad_norm": 15.375,
|
| 242 |
+
"learning_rate": 3.033049640731711e-05,
|
| 243 |
+
"loss": 1.9418,
|
| 244 |
+
"step": 1100
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 0.9777777777777777,
|
| 248 |
+
"eval_loss": 1.9146583080291748,
|
| 249 |
+
"eval_runtime": 48.7562,
|
| 250 |
+
"eval_samples_per_second": 20.51,
|
| 251 |
+
"eval_steps_per_second": 10.255,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.0222222222222221,
|
| 256 |
+
"grad_norm": 11.875,
|
| 257 |
+
"learning_rate": 2.8421706694069926e-05,
|
| 258 |
+
"loss": 1.8229,
|
| 259 |
+
"step": 1150
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.0666666666666667,
|
| 263 |
+
"grad_norm": 13.625,
|
| 264 |
+
"learning_rate": 2.649233843415149e-05,
|
| 265 |
+
"loss": 1.6831,
|
| 266 |
+
"step": 1200
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.0666666666666667,
|
| 270 |
+
"eval_loss": 1.9151355028152466,
|
| 271 |
+
"eval_runtime": 48.9802,
|
| 272 |
+
"eval_samples_per_second": 20.416,
|
| 273 |
+
"eval_steps_per_second": 10.208,
|
| 274 |
+
"step": 1200
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.1111111111111112,
|
| 278 |
+
"grad_norm": 12.625,
|
| 279 |
+
"learning_rate": 2.4553995075294933e-05,
|
| 280 |
+
"loss": 1.6801,
|
| 281 |
+
"step": 1250
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.1555555555555554,
|
| 285 |
+
"grad_norm": 14.5,
|
| 286 |
+
"learning_rate": 2.2618334042534464e-05,
|
| 287 |
+
"loss": 1.7079,
|
| 288 |
+
"step": 1300
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.1555555555555554,
|
| 292 |
+
"eval_loss": 1.9095146656036377,
|
| 293 |
+
"eval_runtime": 48.6949,
|
| 294 |
+
"eval_samples_per_second": 20.536,
|
| 295 |
+
"eval_steps_per_second": 10.268,
|
| 296 |
+
"step": 1300
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.2,
|
| 300 |
+
"grad_norm": 11.75,
|
| 301 |
+
"learning_rate": 2.0696996629079526e-05,
|
| 302 |
+
"loss": 1.6667,
|
| 303 |
+
"step": 1350
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.2444444444444445,
|
| 307 |
+
"grad_norm": 13.375,
|
| 308 |
+
"learning_rate": 1.880153798420768e-05,
|
| 309 |
+
"loss": 1.6615,
|
| 310 |
+
"step": 1400
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.2444444444444445,
|
| 314 |
+
"eval_loss": 1.9029484987258911,
|
| 315 |
+
"eval_runtime": 48.6447,
|
| 316 |
+
"eval_samples_per_second": 20.557,
|
| 317 |
+
"eval_steps_per_second": 10.279,
|
| 318 |
+
"step": 1400
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.2888888888888888,
|
| 322 |
+
"grad_norm": 15.3125,
|
| 323 |
+
"learning_rate": 1.6943357619237226e-05,
|
| 324 |
+
"loss": 1.6417,
|
| 325 |
+
"step": 1450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.3333333333333333,
|
| 329 |
+
"grad_norm": 12.75,
|
| 330 |
+
"learning_rate": 1.5133630849524793e-05,
|
| 331 |
+
"loss": 1.6729,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.3333333333333333,
|
| 336 |
+
"eval_loss": 1.8957328796386719,
|
| 337 |
+
"eval_runtime": 48.7052,
|
| 338 |
+
"eval_samples_per_second": 20.532,
|
| 339 |
+
"eval_steps_per_second": 10.266,
|
| 340 |
+
"step": 1500
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.3777777777777778,
|
| 344 |
+
"grad_norm": 12.8125,
|
| 345 |
+
"learning_rate": 1.3383241584803884e-05,
|
| 346 |
+
"loss": 1.6703,
|
| 347 |
+
"step": 1550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.4222222222222223,
|
| 351 |
+
"grad_norm": 12.3125,
|
| 352 |
+
"learning_rate": 1.170271687207106e-05,
|
| 353 |
+
"loss": 1.6666,
|
| 354 |
+
"step": 1600
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.4222222222222223,
|
| 358 |
+
"eval_loss": 1.89302659034729,
|
| 359 |
+
"eval_runtime": 48.6027,
|
| 360 |
+
"eval_samples_per_second": 20.575,
|
| 361 |
+
"eval_steps_per_second": 10.287,
|
| 362 |
+
"step": 1600
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.4666666666666668,
|
| 366 |
+
"grad_norm": 15.5,
|
| 367 |
+
"learning_rate": 1.010216358468665e-05,
|
| 368 |
+
"loss": 1.6447,
|
| 369 |
+
"step": 1650
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.511111111111111,
|
| 373 |
+
"grad_norm": 12.625,
|
| 374 |
+
"learning_rate": 8.591207638449154e-06,
|
| 375 |
+
"loss": 1.6755,
|
| 376 |
+
"step": 1700
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.511111111111111,
|
| 380 |
+
"eval_loss": 1.891126275062561,
|
| 381 |
+
"eval_runtime": 48.5903,
|
| 382 |
+
"eval_samples_per_second": 20.58,
|
| 383 |
+
"eval_steps_per_second": 10.29,
|
| 384 |
+
"step": 1700
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.5555555555555556,
|
| 388 |
+
"grad_norm": 14.0625,
|
| 389 |
+
"learning_rate": 7.178936100204994e-06,
|
| 390 |
+
"loss": 1.6474,
|
| 391 |
+
"step": 1750
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 1.6,
|
| 395 |
+
"grad_norm": 12.6875,
|
| 396 |
+
"learning_rate": 5.873842537159274e-06,
|
| 397 |
+
"loss": 1.66,
|
| 398 |
+
"step": 1800
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.6,
|
| 402 |
+
"eval_loss": 1.8891392946243286,
|
| 403 |
+
"eval_runtime": 48.6956,
|
| 404 |
+
"eval_samples_per_second": 20.536,
|
| 405 |
+
"eval_steps_per_second": 10.268,
|
| 406 |
+
"step": 1800
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 1.6444444444444444,
|
| 410 |
+
"grad_norm": 14.0625,
|
| 411 |
+
"learning_rate": 4.683775935563222e-06,
|
| 412 |
+
"loss": 1.682,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.6888888888888889,
|
| 417 |
+
"grad_norm": 14.6875,
|
| 418 |
+
"learning_rate": 3.6158934959873353e-06,
|
| 419 |
+
"loss": 1.6299,
|
| 420 |
+
"step": 1900
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.6888888888888889,
|
| 424 |
+
"eval_loss": 1.8877633810043335,
|
| 425 |
+
"eval_runtime": 48.685,
|
| 426 |
+
"eval_samples_per_second": 20.54,
|
| 427 |
+
"eval_steps_per_second": 10.27,
|
| 428 |
+
"step": 1900
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 1.7333333333333334,
|
| 432 |
+
"grad_norm": 11.875,
|
| 433 |
+
"learning_rate": 2.6766175890749786e-06,
|
| 434 |
+
"loss": 1.6601,
|
| 435 |
+
"step": 1950
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.7777777777777777,
|
| 439 |
+
"grad_norm": 13.625,
|
| 440 |
+
"learning_rate": 1.8715971306496745e-06,
|
| 441 |
+
"loss": 1.6286,
|
| 442 |
+
"step": 2000
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.7777777777777777,
|
| 446 |
+
"eval_loss": 1.8876391649246216,
|
| 447 |
+
"eval_runtime": 48.4452,
|
| 448 |
+
"eval_samples_per_second": 20.642,
|
| 449 |
+
"eval_steps_per_second": 10.321,
|
| 450 |
+
"step": 2000
|
| 451 |
+
}
|
| 452 |
+
],
|
| 453 |
+
"logging_steps": 50,
|
| 454 |
+
"max_steps": 2250,
|
| 455 |
+
"num_input_tokens_seen": 0,
|
| 456 |
+
"num_train_epochs": 2,
|
| 457 |
+
"save_steps": 200,
|
| 458 |
+
"stateful_callbacks": {
|
| 459 |
+
"TrainerControl": {
|
| 460 |
+
"args": {
|
| 461 |
+
"should_epoch_stop": false,
|
| 462 |
+
"should_evaluate": false,
|
| 463 |
+
"should_log": false,
|
| 464 |
+
"should_save": true,
|
| 465 |
+
"should_training_stop": false
|
| 466 |
+
},
|
| 467 |
+
"attributes": {}
|
| 468 |
+
}
|
| 469 |
+
},
|
| 470 |
+
"total_flos": 2465612169216000.0,
|
| 471 |
+
"train_batch_size": 2,
|
| 472 |
+
"trial_name": null,
|
| 473 |
+
"trial_params": null
|
| 474 |
+
}
|
checkpoint-2000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51ff9f61bcdaae5ac53531ca1cc31aab96bc0e3a0c0ccd88418c58ca3f44c82d
|
| 3 |
+
size 5304
|
checkpoint-2200/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<image_soft_token>": 262144
|
| 3 |
+
}
|
checkpoint-2200/chat_template.jinja
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{{ bos_token }}
|
| 2 |
+
{%- if messages[0]['role'] == 'system' -%}
|
| 3 |
+
{%- if messages[0]['content'] is string -%}
|
| 4 |
+
{%- set first_user_prefix = messages[0]['content'] + '
|
| 5 |
+
|
| 6 |
+
' -%}
|
| 7 |
+
{%- else -%}
|
| 8 |
+
{%- set first_user_prefix = messages[0]['content'][0]['text'] + '
|
| 9 |
+
|
| 10 |
+
' -%}
|
| 11 |
+
{%- endif -%}
|
| 12 |
+
{%- set loop_messages = messages[1:] -%}
|
| 13 |
+
{%- else -%}
|
| 14 |
+
{%- set first_user_prefix = "" -%}
|
| 15 |
+
{%- set loop_messages = messages -%}
|
| 16 |
+
{%- endif -%}
|
| 17 |
+
{%- for message in loop_messages -%}
|
| 18 |
+
{%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
|
| 19 |
+
{{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
|
| 20 |
+
{%- endif -%}
|
| 21 |
+
{%- if (message['role'] == 'assistant') -%}
|
| 22 |
+
{%- set role = "model" -%}
|
| 23 |
+
{%- else -%}
|
| 24 |
+
{%- set role = message['role'] -%}
|
| 25 |
+
{%- endif -%}
|
| 26 |
+
{{ '<start_of_turn>' + role + '
|
| 27 |
+
' + (first_user_prefix if loop.first else "") }}
|
| 28 |
+
{%- if message['content'] is string -%}
|
| 29 |
+
{{ message['content'] | trim }}
|
| 30 |
+
{%- elif message['content'] is iterable -%}
|
| 31 |
+
{%- for item in message['content'] -%}
|
| 32 |
+
{%- if item['type'] == 'image' -%}
|
| 33 |
+
{{ '<start_of_image>' }}
|
| 34 |
+
{%- elif item['type'] == 'text' -%}
|
| 35 |
+
{{ item['text'] | trim }}
|
| 36 |
+
{%- endif -%}
|
| 37 |
+
{%- endfor -%}
|
| 38 |
+
{%- else -%}
|
| 39 |
+
{{ raise_exception("Invalid content type") }}
|
| 40 |
+
{%- endif -%}
|
| 41 |
+
{{ '<end_of_turn>
|
| 42 |
+
' }}
|
| 43 |
+
{%- endfor -%}
|
| 44 |
+
{%- if add_generation_prompt -%}
|
| 45 |
+
{{'<start_of_turn>model
|
| 46 |
+
'}}
|
| 47 |
+
{%- endif -%}
|
checkpoint-2200/config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_sliding_window_pattern": 6,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Gemma3ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_logit_softcapping": null,
|
| 9 |
+
"bos_token_id": 2,
|
| 10 |
+
"cache_implementation": "hybrid",
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"final_logit_softcapping": null,
|
| 13 |
+
"head_dim": 256,
|
| 14 |
+
"hidden_activation": "gelu_pytorch_tanh",
|
| 15 |
+
"hidden_size": 640,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 2048,
|
| 18 |
+
"layer_types": [
|
| 19 |
+
"sliding_attention",
|
| 20 |
+
"sliding_attention",
|
| 21 |
+
"sliding_attention",
|
| 22 |
+
"sliding_attention",
|
| 23 |
+
"sliding_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"sliding_attention",
|
| 26 |
+
"sliding_attention",
|
| 27 |
+
"sliding_attention",
|
| 28 |
+
"sliding_attention",
|
| 29 |
+
"sliding_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"sliding_attention",
|
| 32 |
+
"sliding_attention",
|
| 33 |
+
"sliding_attention",
|
| 34 |
+
"sliding_attention",
|
| 35 |
+
"sliding_attention",
|
| 36 |
+
"full_attention"
|
| 37 |
+
],
|
| 38 |
+
"max_position_embeddings": 32768,
|
| 39 |
+
"model_type": "gemma3_text",
|
| 40 |
+
"num_attention_heads": 4,
|
| 41 |
+
"num_hidden_layers": 18,
|
| 42 |
+
"num_key_value_heads": 1,
|
| 43 |
+
"pad_token_id": 0,
|
| 44 |
+
"query_pre_attn_scalar": 256,
|
| 45 |
+
"rms_norm_eps": 1e-06,
|
| 46 |
+
"rope_local_base_freq": 10000.0,
|
| 47 |
+
"rope_scaling": null,
|
| 48 |
+
"rope_theta": 1000000.0,
|
| 49 |
+
"sliding_window": 512,
|
| 50 |
+
"sliding_window_pattern": 6,
|
| 51 |
+
"torch_dtype": "bfloat16",
|
| 52 |
+
"transformers_version": "4.52.4",
|
| 53 |
+
"use_bidirectional_attention": false,
|
| 54 |
+
"use_cache": true,
|
| 55 |
+
"vocab_size": 262144
|
| 56 |
+
}
|
checkpoint-2200/generation_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cache_implementation": "hybrid",
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
1,
|
| 6 |
+
106
|
| 7 |
+
],
|
| 8 |
+
"top_k": 64,
|
| 9 |
+
"top_p": 0.95,
|
| 10 |
+
"transformers_version": "4.52.4"
|
| 11 |
+
}
|
checkpoint-2200/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:efe872bec3b6dd893add9a14c1f41c7221af35506062d825c0c88f42f9065b80
|
| 3 |
+
size 536223056
|
checkpoint-2200/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33f9ef975b278600b0268b01644e66932af546a098a8007c4900562970239155
|
| 3 |
+
size 1072590714
|
checkpoint-2200/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5f38f6446ba43bb9ae0be4911c150b41c5adfbe2712844ea3cb854b62ee2432
|
| 3 |
+
size 14244
|
checkpoint-2200/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:78b6b554e381cf0f46cecfb4e224f04cd835ca288ef75eb3df2c929e21a0cb8a
|
| 3 |
+
size 1064
|
checkpoint-2200/special_tokens_map.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"boi_token": "<start_of_image>",
|
| 3 |
+
"bos_token": {
|
| 4 |
+
"content": "<bos>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"eoi_token": "<end_of_image>",
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"content": "<eos>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false
|
| 17 |
+
},
|
| 18 |
+
"image_token": "<image_soft_token>",
|
| 19 |
+
"pad_token": {
|
| 20 |
+
"content": "<pad>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false
|
| 25 |
+
},
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": false,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
}
|
| 33 |
+
}
|
checkpoint-2200/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e570d6288ff0afcea981a80492eddfa3e2239a79de89e5074cbb74b548fa5e2b
|
| 3 |
+
size 33384833
|
checkpoint-2200/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
|
| 3 |
+
size 4689074
|
checkpoint-2200/tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-2200/trainer_state.json
ADDED
|
@@ -0,0 +1,518 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2000,
|
| 3 |
+
"best_metric": 1.8876391649246216,
|
| 4 |
+
"best_model_checkpoint": "./gemma3-270m-turkish_instructions-finetuned/checkpoint-2000",
|
| 5 |
+
"epoch": 1.9555555555555557,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 2200,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.044444444444444446,
|
| 14 |
+
"grad_norm": 19.25,
|
| 15 |
+
"learning_rate": 1.088888888888889e-05,
|
| 16 |
+
"loss": 3.5828,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.08888888888888889,
|
| 21 |
+
"grad_norm": 17.875,
|
| 22 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 23 |
+
"loss": 2.4877,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.08888888888888889,
|
| 28 |
+
"eval_loss": 2.275900363922119,
|
| 29 |
+
"eval_runtime": 48.3609,
|
| 30 |
+
"eval_samples_per_second": 20.678,
|
| 31 |
+
"eval_steps_per_second": 10.339,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.13333333333333333,
|
| 36 |
+
"grad_norm": 18.25,
|
| 37 |
+
"learning_rate": 3.311111111111112e-05,
|
| 38 |
+
"loss": 2.255,
|
| 39 |
+
"step": 150
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.17777777777777778,
|
| 43 |
+
"grad_norm": 13.4375,
|
| 44 |
+
"learning_rate": 4.422222222222222e-05,
|
| 45 |
+
"loss": 2.2559,
|
| 46 |
+
"step": 200
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.17777777777777778,
|
| 50 |
+
"eval_loss": 2.1721303462982178,
|
| 51 |
+
"eval_runtime": 48.5933,
|
| 52 |
+
"eval_samples_per_second": 20.579,
|
| 53 |
+
"eval_steps_per_second": 10.289,
|
| 54 |
+
"step": 200
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.2222222222222222,
|
| 58 |
+
"grad_norm": 18.875,
|
| 59 |
+
"learning_rate": 4.99826726554013e-05,
|
| 60 |
+
"loss": 2.2218,
|
| 61 |
+
"step": 250
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.26666666666666666,
|
| 65 |
+
"grad_norm": 16.625,
|
| 66 |
+
"learning_rate": 4.983543173414964e-05,
|
| 67 |
+
"loss": 2.1805,
|
| 68 |
+
"step": 300
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26666666666666666,
|
| 72 |
+
"eval_loss": 2.1423935890197754,
|
| 73 |
+
"eval_runtime": 48.7813,
|
| 74 |
+
"eval_samples_per_second": 20.5,
|
| 75 |
+
"eval_steps_per_second": 10.25,
|
| 76 |
+
"step": 300
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.3111111111111111,
|
| 80 |
+
"grad_norm": 15.0625,
|
| 81 |
+
"learning_rate": 4.953882760420223e-05,
|
| 82 |
+
"loss": 2.2345,
|
| 83 |
+
"step": 350
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.35555555555555557,
|
| 87 |
+
"grad_norm": 15.75,
|
| 88 |
+
"learning_rate": 4.909464407769633e-05,
|
| 89 |
+
"loss": 2.1545,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.35555555555555557,
|
| 94 |
+
"eval_loss": 2.081547260284424,
|
| 95 |
+
"eval_runtime": 48.8759,
|
| 96 |
+
"eval_samples_per_second": 20.46,
|
| 97 |
+
"eval_steps_per_second": 10.23,
|
| 98 |
+
"step": 400
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.4,
|
| 102 |
+
"grad_norm": 17.875,
|
| 103 |
+
"learning_rate": 4.850555252662495e-05,
|
| 104 |
+
"loss": 2.0966,
|
| 105 |
+
"step": 450
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.4444444444444444,
|
| 109 |
+
"grad_norm": 13.1875,
|
| 110 |
+
"learning_rate": 4.7775095816891336e-05,
|
| 111 |
+
"loss": 2.0782,
|
| 112 |
+
"step": 500
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.4444444444444444,
|
| 116 |
+
"eval_loss": 2.0508856773376465,
|
| 117 |
+
"eval_runtime": 48.9079,
|
| 118 |
+
"eval_samples_per_second": 20.447,
|
| 119 |
+
"eval_steps_per_second": 10.223,
|
| 120 |
+
"step": 500
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.4888888888888889,
|
| 124 |
+
"grad_norm": 15.9375,
|
| 125 |
+
"learning_rate": 4.690766700109659e-05,
|
| 126 |
+
"loss": 2.115,
|
| 127 |
+
"step": 550
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.5333333333333333,
|
| 131 |
+
"grad_norm": 13.9375,
|
| 132 |
+
"learning_rate": 4.590848289820442e-05,
|
| 133 |
+
"loss": 2.0314,
|
| 134 |
+
"step": 600
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.5333333333333333,
|
| 138 |
+
"eval_loss": 2.0262327194213867,
|
| 139 |
+
"eval_runtime": 48.768,
|
| 140 |
+
"eval_samples_per_second": 20.505,
|
| 141 |
+
"eval_steps_per_second": 10.253,
|
| 142 |
+
"step": 600
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.5777777777777777,
|
| 146 |
+
"grad_norm": 14.5,
|
| 147 |
+
"learning_rate": 4.4783552718978e-05,
|
| 148 |
+
"loss": 2.0892,
|
| 149 |
+
"step": 650
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.6222222222222222,
|
| 153 |
+
"grad_norm": 14.0,
|
| 154 |
+
"learning_rate": 4.3539641925879495e-05,
|
| 155 |
+
"loss": 2.0352,
|
| 156 |
+
"step": 700
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.6222222222222222,
|
| 160 |
+
"eval_loss": 1.9995155334472656,
|
| 161 |
+
"eval_runtime": 48.8752,
|
| 162 |
+
"eval_samples_per_second": 20.46,
|
| 163 |
+
"eval_steps_per_second": 10.23,
|
| 164 |
+
"step": 700
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.6666666666666666,
|
| 168 |
+
"grad_norm": 13.25,
|
| 169 |
+
"learning_rate": 4.2184231544782596e-05,
|
| 170 |
+
"loss": 2.0117,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.7111111111111111,
|
| 175 |
+
"grad_norm": 13.25,
|
| 176 |
+
"learning_rate": 4.072547317320281e-05,
|
| 177 |
+
"loss": 1.9848,
|
| 178 |
+
"step": 800
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.7111111111111111,
|
| 182 |
+
"eval_loss": 1.9770301580429077,
|
| 183 |
+
"eval_runtime": 48.7445,
|
| 184 |
+
"eval_samples_per_second": 20.515,
|
| 185 |
+
"eval_steps_per_second": 10.258,
|
| 186 |
+
"step": 800
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.7555555555555555,
|
| 190 |
+
"grad_norm": 13.375,
|
| 191 |
+
"learning_rate": 3.9172139955630774e-05,
|
| 192 |
+
"loss": 2.0014,
|
| 193 |
+
"step": 850
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.8,
|
| 197 |
+
"grad_norm": 12.75,
|
| 198 |
+
"learning_rate": 3.7533573820809006e-05,
|
| 199 |
+
"loss": 1.9938,
|
| 200 |
+
"step": 900
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.8,
|
| 204 |
+
"eval_loss": 1.9461404085159302,
|
| 205 |
+
"eval_runtime": 48.8094,
|
| 206 |
+
"eval_samples_per_second": 20.488,
|
| 207 |
+
"eval_steps_per_second": 10.244,
|
| 208 |
+
"step": 900
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.8444444444444444,
|
| 212 |
+
"grad_norm": 13.8125,
|
| 213 |
+
"learning_rate": 3.5819629298273245e-05,
|
| 214 |
+
"loss": 1.9914,
|
| 215 |
+
"step": 950
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.8888888888888888,
|
| 219 |
+
"grad_norm": 13.6875,
|
| 220 |
+
"learning_rate": 3.4040614252052305e-05,
|
| 221 |
+
"loss": 1.9724,
|
| 222 |
+
"step": 1000
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.8888888888888888,
|
| 226 |
+
"eval_loss": 1.9275351762771606,
|
| 227 |
+
"eval_runtime": 48.8023,
|
| 228 |
+
"eval_samples_per_second": 20.491,
|
| 229 |
+
"eval_steps_per_second": 10.245,
|
| 230 |
+
"step": 1000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.9333333333333333,
|
| 234 |
+
"grad_norm": 13.25,
|
| 235 |
+
"learning_rate": 3.2207227887960935e-05,
|
| 236 |
+
"loss": 1.9607,
|
| 237 |
+
"step": 1050
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.9777777777777777,
|
| 241 |
+
"grad_norm": 15.375,
|
| 242 |
+
"learning_rate": 3.033049640731711e-05,
|
| 243 |
+
"loss": 1.9418,
|
| 244 |
+
"step": 1100
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 0.9777777777777777,
|
| 248 |
+
"eval_loss": 1.9146583080291748,
|
| 249 |
+
"eval_runtime": 48.7562,
|
| 250 |
+
"eval_samples_per_second": 20.51,
|
| 251 |
+
"eval_steps_per_second": 10.255,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.0222222222222221,
|
| 256 |
+
"grad_norm": 11.875,
|
| 257 |
+
"learning_rate": 2.8421706694069926e-05,
|
| 258 |
+
"loss": 1.8229,
|
| 259 |
+
"step": 1150
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.0666666666666667,
|
| 263 |
+
"grad_norm": 13.625,
|
| 264 |
+
"learning_rate": 2.649233843415149e-05,
|
| 265 |
+
"loss": 1.6831,
|
| 266 |
+
"step": 1200
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.0666666666666667,
|
| 270 |
+
"eval_loss": 1.9151355028152466,
|
| 271 |
+
"eval_runtime": 48.9802,
|
| 272 |
+
"eval_samples_per_second": 20.416,
|
| 273 |
+
"eval_steps_per_second": 10.208,
|
| 274 |
+
"step": 1200
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.1111111111111112,
|
| 278 |
+
"grad_norm": 12.625,
|
| 279 |
+
"learning_rate": 2.4553995075294933e-05,
|
| 280 |
+
"loss": 1.6801,
|
| 281 |
+
"step": 1250
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.1555555555555554,
|
| 285 |
+
"grad_norm": 14.5,
|
| 286 |
+
"learning_rate": 2.2618334042534464e-05,
|
| 287 |
+
"loss": 1.7079,
|
| 288 |
+
"step": 1300
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.1555555555555554,
|
| 292 |
+
"eval_loss": 1.9095146656036377,
|
| 293 |
+
"eval_runtime": 48.6949,
|
| 294 |
+
"eval_samples_per_second": 20.536,
|
| 295 |
+
"eval_steps_per_second": 10.268,
|
| 296 |
+
"step": 1300
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.2,
|
| 300 |
+
"grad_norm": 11.75,
|
| 301 |
+
"learning_rate": 2.0696996629079526e-05,
|
| 302 |
+
"loss": 1.6667,
|
| 303 |
+
"step": 1350
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.2444444444444445,
|
| 307 |
+
"grad_norm": 13.375,
|
| 308 |
+
"learning_rate": 1.880153798420768e-05,
|
| 309 |
+
"loss": 1.6615,
|
| 310 |
+
"step": 1400
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.2444444444444445,
|
| 314 |
+
"eval_loss": 1.9029484987258911,
|
| 315 |
+
"eval_runtime": 48.6447,
|
| 316 |
+
"eval_samples_per_second": 20.557,
|
| 317 |
+
"eval_steps_per_second": 10.279,
|
| 318 |
+
"step": 1400
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.2888888888888888,
|
| 322 |
+
"grad_norm": 15.3125,
|
| 323 |
+
"learning_rate": 1.6943357619237226e-05,
|
| 324 |
+
"loss": 1.6417,
|
| 325 |
+
"step": 1450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.3333333333333333,
|
| 329 |
+
"grad_norm": 12.75,
|
| 330 |
+
"learning_rate": 1.5133630849524793e-05,
|
| 331 |
+
"loss": 1.6729,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.3333333333333333,
|
| 336 |
+
"eval_loss": 1.8957328796386719,
|
| 337 |
+
"eval_runtime": 48.7052,
|
| 338 |
+
"eval_samples_per_second": 20.532,
|
| 339 |
+
"eval_steps_per_second": 10.266,
|
| 340 |
+
"step": 1500
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.3777777777777778,
|
| 344 |
+
"grad_norm": 12.8125,
|
| 345 |
+
"learning_rate": 1.3383241584803884e-05,
|
| 346 |
+
"loss": 1.6703,
|
| 347 |
+
"step": 1550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.4222222222222223,
|
| 351 |
+
"grad_norm": 12.3125,
|
| 352 |
+
"learning_rate": 1.170271687207106e-05,
|
| 353 |
+
"loss": 1.6666,
|
| 354 |
+
"step": 1600
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.4222222222222223,
|
| 358 |
+
"eval_loss": 1.89302659034729,
|
| 359 |
+
"eval_runtime": 48.6027,
|
| 360 |
+
"eval_samples_per_second": 20.575,
|
| 361 |
+
"eval_steps_per_second": 10.287,
|
| 362 |
+
"step": 1600
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.4666666666666668,
|
| 366 |
+
"grad_norm": 15.5,
|
| 367 |
+
"learning_rate": 1.010216358468665e-05,
|
| 368 |
+
"loss": 1.6447,
|
| 369 |
+
"step": 1650
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.511111111111111,
|
| 373 |
+
"grad_norm": 12.625,
|
| 374 |
+
"learning_rate": 8.591207638449154e-06,
|
| 375 |
+
"loss": 1.6755,
|
| 376 |
+
"step": 1700
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.511111111111111,
|
| 380 |
+
"eval_loss": 1.891126275062561,
|
| 381 |
+
"eval_runtime": 48.5903,
|
| 382 |
+
"eval_samples_per_second": 20.58,
|
| 383 |
+
"eval_steps_per_second": 10.29,
|
| 384 |
+
"step": 1700
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.5555555555555556,
|
| 388 |
+
"grad_norm": 14.0625,
|
| 389 |
+
"learning_rate": 7.178936100204994e-06,
|
| 390 |
+
"loss": 1.6474,
|
| 391 |
+
"step": 1750
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 1.6,
|
| 395 |
+
"grad_norm": 12.6875,
|
| 396 |
+
"learning_rate": 5.873842537159274e-06,
|
| 397 |
+
"loss": 1.66,
|
| 398 |
+
"step": 1800
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.6,
|
| 402 |
+
"eval_loss": 1.8891392946243286,
|
| 403 |
+
"eval_runtime": 48.6956,
|
| 404 |
+
"eval_samples_per_second": 20.536,
|
| 405 |
+
"eval_steps_per_second": 10.268,
|
| 406 |
+
"step": 1800
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 1.6444444444444444,
|
| 410 |
+
"grad_norm": 14.0625,
|
| 411 |
+
"learning_rate": 4.683775935563222e-06,
|
| 412 |
+
"loss": 1.682,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.6888888888888889,
|
| 417 |
+
"grad_norm": 14.6875,
|
| 418 |
+
"learning_rate": 3.6158934959873353e-06,
|
| 419 |
+
"loss": 1.6299,
|
| 420 |
+
"step": 1900
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.6888888888888889,
|
| 424 |
+
"eval_loss": 1.8877633810043335,
|
| 425 |
+
"eval_runtime": 48.685,
|
| 426 |
+
"eval_samples_per_second": 20.54,
|
| 427 |
+
"eval_steps_per_second": 10.27,
|
| 428 |
+
"step": 1900
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 1.7333333333333334,
|
| 432 |
+
"grad_norm": 11.875,
|
| 433 |
+
"learning_rate": 2.6766175890749786e-06,
|
| 434 |
+
"loss": 1.6601,
|
| 435 |
+
"step": 1950
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.7777777777777777,
|
| 439 |
+
"grad_norm": 13.625,
|
| 440 |
+
"learning_rate": 1.8715971306496745e-06,
|
| 441 |
+
"loss": 1.6286,
|
| 442 |
+
"step": 2000
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.7777777777777777,
|
| 446 |
+
"eval_loss": 1.8876391649246216,
|
| 447 |
+
"eval_runtime": 48.4452,
|
| 448 |
+
"eval_samples_per_second": 20.642,
|
| 449 |
+
"eval_steps_per_second": 10.321,
|
| 450 |
+
"step": 2000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.8222222222222222,
|
| 454 |
+
"grad_norm": 11.625,
|
| 455 |
+
"learning_rate": 1.2056736084706589e-06,
|
| 456 |
+
"loss": 1.6788,
|
| 457 |
+
"step": 2050
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.8666666666666667,
|
| 461 |
+
"grad_norm": 13.125,
|
| 462 |
+
"learning_rate": 6.828519649558191e-07,
|
| 463 |
+
"loss": 1.6811,
|
| 464 |
+
"step": 2100
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.8666666666666667,
|
| 468 |
+
"eval_loss": 1.8877239227294922,
|
| 469 |
+
"eval_runtime": 48.6643,
|
| 470 |
+
"eval_samples_per_second": 20.549,
|
| 471 |
+
"eval_steps_per_second": 10.274,
|
| 472 |
+
"step": 2100
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.911111111111111,
|
| 476 |
+
"grad_norm": 12.1875,
|
| 477 |
+
"learning_rate": 3.062765109867499e-07,
|
| 478 |
+
"loss": 1.6562,
|
| 479 |
+
"step": 2150
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.9555555555555557,
|
| 483 |
+
"grad_norm": 13.4375,
|
| 484 |
+
"learning_rate": 7.821201565316182e-08,
|
| 485 |
+
"loss": 1.6527,
|
| 486 |
+
"step": 2200
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.9555555555555557,
|
| 490 |
+
"eval_loss": 1.8878294229507446,
|
| 491 |
+
"eval_runtime": 48.65,
|
| 492 |
+
"eval_samples_per_second": 20.555,
|
| 493 |
+
"eval_steps_per_second": 10.277,
|
| 494 |
+
"step": 2200
|
| 495 |
+
}
|
| 496 |
+
],
|
| 497 |
+
"logging_steps": 50,
|
| 498 |
+
"max_steps": 2250,
|
| 499 |
+
"num_input_tokens_seen": 0,
|
| 500 |
+
"num_train_epochs": 2,
|
| 501 |
+
"save_steps": 200,
|
| 502 |
+
"stateful_callbacks": {
|
| 503 |
+
"TrainerControl": {
|
| 504 |
+
"args": {
|
| 505 |
+
"should_epoch_stop": false,
|
| 506 |
+
"should_evaluate": false,
|
| 507 |
+
"should_log": false,
|
| 508 |
+
"should_save": true,
|
| 509 |
+
"should_training_stop": false
|
| 510 |
+
},
|
| 511 |
+
"attributes": {}
|
| 512 |
+
}
|
| 513 |
+
},
|
| 514 |
+
"total_flos": 2712173386137600.0,
|
| 515 |
+
"train_batch_size": 2,
|
| 516 |
+
"trial_name": null,
|
| 517 |
+
"trial_params": null
|
| 518 |
+
}
|
checkpoint-2200/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51ff9f61bcdaae5ac53531ca1cc31aab96bc0e3a0c0ccd88418c58ca3f44c82d
|
| 3 |
+
size 5304
|
checkpoint-2250/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<image_soft_token>": 262144
|
| 3 |
+
}
|
checkpoint-2250/chat_template.jinja
ADDED
|
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{{ bos_token }}
|
| 2 |
+
{%- if messages[0]['role'] == 'system' -%}
|
| 3 |
+
{%- if messages[0]['content'] is string -%}
|
| 4 |
+
{%- set first_user_prefix = messages[0]['content'] + '
|
| 5 |
+
|
| 6 |
+
' -%}
|
| 7 |
+
{%- else -%}
|
| 8 |
+
{%- set first_user_prefix = messages[0]['content'][0]['text'] + '
|
| 9 |
+
|
| 10 |
+
' -%}
|
| 11 |
+
{%- endif -%}
|
| 12 |
+
{%- set loop_messages = messages[1:] -%}
|
| 13 |
+
{%- else -%}
|
| 14 |
+
{%- set first_user_prefix = "" -%}
|
| 15 |
+
{%- set loop_messages = messages -%}
|
| 16 |
+
{%- endif -%}
|
| 17 |
+
{%- for message in loop_messages -%}
|
| 18 |
+
{%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}
|
| 19 |
+
{{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
|
| 20 |
+
{%- endif -%}
|
| 21 |
+
{%- if (message['role'] == 'assistant') -%}
|
| 22 |
+
{%- set role = "model" -%}
|
| 23 |
+
{%- else -%}
|
| 24 |
+
{%- set role = message['role'] -%}
|
| 25 |
+
{%- endif -%}
|
| 26 |
+
{{ '<start_of_turn>' + role + '
|
| 27 |
+
' + (first_user_prefix if loop.first else "") }}
|
| 28 |
+
{%- if message['content'] is string -%}
|
| 29 |
+
{{ message['content'] | trim }}
|
| 30 |
+
{%- elif message['content'] is iterable -%}
|
| 31 |
+
{%- for item in message['content'] -%}
|
| 32 |
+
{%- if item['type'] == 'image' -%}
|
| 33 |
+
{{ '<start_of_image>' }}
|
| 34 |
+
{%- elif item['type'] == 'text' -%}
|
| 35 |
+
{{ item['text'] | trim }}
|
| 36 |
+
{%- endif -%}
|
| 37 |
+
{%- endfor -%}
|
| 38 |
+
{%- else -%}
|
| 39 |
+
{{ raise_exception("Invalid content type") }}
|
| 40 |
+
{%- endif -%}
|
| 41 |
+
{{ '<end_of_turn>
|
| 42 |
+
' }}
|
| 43 |
+
{%- endfor -%}
|
| 44 |
+
{%- if add_generation_prompt -%}
|
| 45 |
+
{{'<start_of_turn>model
|
| 46 |
+
'}}
|
| 47 |
+
{%- endif -%}
|
checkpoint-2250/config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_sliding_window_pattern": 6,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Gemma3ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_logit_softcapping": null,
|
| 9 |
+
"bos_token_id": 2,
|
| 10 |
+
"cache_implementation": "hybrid",
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"final_logit_softcapping": null,
|
| 13 |
+
"head_dim": 256,
|
| 14 |
+
"hidden_activation": "gelu_pytorch_tanh",
|
| 15 |
+
"hidden_size": 640,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 2048,
|
| 18 |
+
"layer_types": [
|
| 19 |
+
"sliding_attention",
|
| 20 |
+
"sliding_attention",
|
| 21 |
+
"sliding_attention",
|
| 22 |
+
"sliding_attention",
|
| 23 |
+
"sliding_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"sliding_attention",
|
| 26 |
+
"sliding_attention",
|
| 27 |
+
"sliding_attention",
|
| 28 |
+
"sliding_attention",
|
| 29 |
+
"sliding_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"sliding_attention",
|
| 32 |
+
"sliding_attention",
|
| 33 |
+
"sliding_attention",
|
| 34 |
+
"sliding_attention",
|
| 35 |
+
"sliding_attention",
|
| 36 |
+
"full_attention"
|
| 37 |
+
],
|
| 38 |
+
"max_position_embeddings": 32768,
|
| 39 |
+
"model_type": "gemma3_text",
|
| 40 |
+
"num_attention_heads": 4,
|
| 41 |
+
"num_hidden_layers": 18,
|
| 42 |
+
"num_key_value_heads": 1,
|
| 43 |
+
"pad_token_id": 0,
|
| 44 |
+
"query_pre_attn_scalar": 256,
|
| 45 |
+
"rms_norm_eps": 1e-06,
|
| 46 |
+
"rope_local_base_freq": 10000.0,
|
| 47 |
+
"rope_scaling": null,
|
| 48 |
+
"rope_theta": 1000000.0,
|
| 49 |
+
"sliding_window": 512,
|
| 50 |
+
"sliding_window_pattern": 6,
|
| 51 |
+
"torch_dtype": "bfloat16",
|
| 52 |
+
"transformers_version": "4.52.4",
|
| 53 |
+
"use_bidirectional_attention": false,
|
| 54 |
+
"use_cache": true,
|
| 55 |
+
"vocab_size": 262144
|
| 56 |
+
}
|
checkpoint-2250/generation_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cache_implementation": "hybrid",
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
1,
|
| 6 |
+
106
|
| 7 |
+
],
|
| 8 |
+
"top_k": 64,
|
| 9 |
+
"top_p": 0.95,
|
| 10 |
+
"transformers_version": "4.52.4"
|
| 11 |
+
}
|
checkpoint-2250/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:167a1b553a36c55710b27020272e1c73fa083db459266e49447d4a5f9fa0e99e
|
| 3 |
+
size 536223056
|
checkpoint-2250/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7ca38aa29fdc7d7fae28324f62f22d3b66a9327f3d1569d34cf169fd82dfd9d
|
| 3 |
+
size 1072590714
|
checkpoint-2250/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d5f38f6446ba43bb9ae0be4911c150b41c5adfbe2712844ea3cb854b62ee2432
|
| 3 |
+
size 14244
|
checkpoint-2250/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1bec953c174e41850d61f3908d88b9683a2b8a87c5e9752e8ae44ce839aa2004
|
| 3 |
+
size 1064
|
checkpoint-2250/special_tokens_map.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"boi_token": "<start_of_image>",
|
| 3 |
+
"bos_token": {
|
| 4 |
+
"content": "<bos>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"eoi_token": "<end_of_image>",
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"content": "<eos>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false
|
| 17 |
+
},
|
| 18 |
+
"image_token": "<image_soft_token>",
|
| 19 |
+
"pad_token": {
|
| 20 |
+
"content": "<pad>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false
|
| 25 |
+
},
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": false,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
}
|
| 33 |
+
}
|
checkpoint-2250/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e570d6288ff0afcea981a80492eddfa3e2239a79de89e5074cbb74b548fa5e2b
|
| 3 |
+
size 33384833
|
checkpoint-2250/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
|
| 3 |
+
size 4689074
|
checkpoint-2250/tokenizer_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-2250/trainer_state.json
ADDED
|
@@ -0,0 +1,525 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2000,
|
| 3 |
+
"best_metric": 1.8876391649246216,
|
| 4 |
+
"best_model_checkpoint": "./gemma3-270m-turkish_instructions-finetuned/checkpoint-2000",
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 2250,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.044444444444444446,
|
| 14 |
+
"grad_norm": 19.25,
|
| 15 |
+
"learning_rate": 1.088888888888889e-05,
|
| 16 |
+
"loss": 3.5828,
|
| 17 |
+
"step": 50
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.08888888888888889,
|
| 21 |
+
"grad_norm": 17.875,
|
| 22 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 23 |
+
"loss": 2.4877,
|
| 24 |
+
"step": 100
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.08888888888888889,
|
| 28 |
+
"eval_loss": 2.275900363922119,
|
| 29 |
+
"eval_runtime": 48.3609,
|
| 30 |
+
"eval_samples_per_second": 20.678,
|
| 31 |
+
"eval_steps_per_second": 10.339,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.13333333333333333,
|
| 36 |
+
"grad_norm": 18.25,
|
| 37 |
+
"learning_rate": 3.311111111111112e-05,
|
| 38 |
+
"loss": 2.255,
|
| 39 |
+
"step": 150
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.17777777777777778,
|
| 43 |
+
"grad_norm": 13.4375,
|
| 44 |
+
"learning_rate": 4.422222222222222e-05,
|
| 45 |
+
"loss": 2.2559,
|
| 46 |
+
"step": 200
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"epoch": 0.17777777777777778,
|
| 50 |
+
"eval_loss": 2.1721303462982178,
|
| 51 |
+
"eval_runtime": 48.5933,
|
| 52 |
+
"eval_samples_per_second": 20.579,
|
| 53 |
+
"eval_steps_per_second": 10.289,
|
| 54 |
+
"step": 200
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.2222222222222222,
|
| 58 |
+
"grad_norm": 18.875,
|
| 59 |
+
"learning_rate": 4.99826726554013e-05,
|
| 60 |
+
"loss": 2.2218,
|
| 61 |
+
"step": 250
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.26666666666666666,
|
| 65 |
+
"grad_norm": 16.625,
|
| 66 |
+
"learning_rate": 4.983543173414964e-05,
|
| 67 |
+
"loss": 2.1805,
|
| 68 |
+
"step": 300
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26666666666666666,
|
| 72 |
+
"eval_loss": 2.1423935890197754,
|
| 73 |
+
"eval_runtime": 48.7813,
|
| 74 |
+
"eval_samples_per_second": 20.5,
|
| 75 |
+
"eval_steps_per_second": 10.25,
|
| 76 |
+
"step": 300
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 0.3111111111111111,
|
| 80 |
+
"grad_norm": 15.0625,
|
| 81 |
+
"learning_rate": 4.953882760420223e-05,
|
| 82 |
+
"loss": 2.2345,
|
| 83 |
+
"step": 350
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.35555555555555557,
|
| 87 |
+
"grad_norm": 15.75,
|
| 88 |
+
"learning_rate": 4.909464407769633e-05,
|
| 89 |
+
"loss": 2.1545,
|
| 90 |
+
"step": 400
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.35555555555555557,
|
| 94 |
+
"eval_loss": 2.081547260284424,
|
| 95 |
+
"eval_runtime": 48.8759,
|
| 96 |
+
"eval_samples_per_second": 20.46,
|
| 97 |
+
"eval_steps_per_second": 10.23,
|
| 98 |
+
"step": 400
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.4,
|
| 102 |
+
"grad_norm": 17.875,
|
| 103 |
+
"learning_rate": 4.850555252662495e-05,
|
| 104 |
+
"loss": 2.0966,
|
| 105 |
+
"step": 450
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.4444444444444444,
|
| 109 |
+
"grad_norm": 13.1875,
|
| 110 |
+
"learning_rate": 4.7775095816891336e-05,
|
| 111 |
+
"loss": 2.0782,
|
| 112 |
+
"step": 500
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.4444444444444444,
|
| 116 |
+
"eval_loss": 2.0508856773376465,
|
| 117 |
+
"eval_runtime": 48.9079,
|
| 118 |
+
"eval_samples_per_second": 20.447,
|
| 119 |
+
"eval_steps_per_second": 10.223,
|
| 120 |
+
"step": 500
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.4888888888888889,
|
| 124 |
+
"grad_norm": 15.9375,
|
| 125 |
+
"learning_rate": 4.690766700109659e-05,
|
| 126 |
+
"loss": 2.115,
|
| 127 |
+
"step": 550
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 0.5333333333333333,
|
| 131 |
+
"grad_norm": 13.9375,
|
| 132 |
+
"learning_rate": 4.590848289820442e-05,
|
| 133 |
+
"loss": 2.0314,
|
| 134 |
+
"step": 600
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.5333333333333333,
|
| 138 |
+
"eval_loss": 2.0262327194213867,
|
| 139 |
+
"eval_runtime": 48.768,
|
| 140 |
+
"eval_samples_per_second": 20.505,
|
| 141 |
+
"eval_steps_per_second": 10.253,
|
| 142 |
+
"step": 600
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.5777777777777777,
|
| 146 |
+
"grad_norm": 14.5,
|
| 147 |
+
"learning_rate": 4.4783552718978e-05,
|
| 148 |
+
"loss": 2.0892,
|
| 149 |
+
"step": 650
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.6222222222222222,
|
| 153 |
+
"grad_norm": 14.0,
|
| 154 |
+
"learning_rate": 4.3539641925879495e-05,
|
| 155 |
+
"loss": 2.0352,
|
| 156 |
+
"step": 700
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.6222222222222222,
|
| 160 |
+
"eval_loss": 1.9995155334472656,
|
| 161 |
+
"eval_runtime": 48.8752,
|
| 162 |
+
"eval_samples_per_second": 20.46,
|
| 163 |
+
"eval_steps_per_second": 10.23,
|
| 164 |
+
"step": 700
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.6666666666666666,
|
| 168 |
+
"grad_norm": 13.25,
|
| 169 |
+
"learning_rate": 4.2184231544782596e-05,
|
| 170 |
+
"loss": 2.0117,
|
| 171 |
+
"step": 750
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.7111111111111111,
|
| 175 |
+
"grad_norm": 13.25,
|
| 176 |
+
"learning_rate": 4.072547317320281e-05,
|
| 177 |
+
"loss": 1.9848,
|
| 178 |
+
"step": 800
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.7111111111111111,
|
| 182 |
+
"eval_loss": 1.9770301580429077,
|
| 183 |
+
"eval_runtime": 48.7445,
|
| 184 |
+
"eval_samples_per_second": 20.515,
|
| 185 |
+
"eval_steps_per_second": 10.258,
|
| 186 |
+
"step": 800
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.7555555555555555,
|
| 190 |
+
"grad_norm": 13.375,
|
| 191 |
+
"learning_rate": 3.9172139955630774e-05,
|
| 192 |
+
"loss": 2.0014,
|
| 193 |
+
"step": 850
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.8,
|
| 197 |
+
"grad_norm": 12.75,
|
| 198 |
+
"learning_rate": 3.7533573820809006e-05,
|
| 199 |
+
"loss": 1.9938,
|
| 200 |
+
"step": 900
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.8,
|
| 204 |
+
"eval_loss": 1.9461404085159302,
|
| 205 |
+
"eval_runtime": 48.8094,
|
| 206 |
+
"eval_samples_per_second": 20.488,
|
| 207 |
+
"eval_steps_per_second": 10.244,
|
| 208 |
+
"step": 900
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.8444444444444444,
|
| 212 |
+
"grad_norm": 13.8125,
|
| 213 |
+
"learning_rate": 3.5819629298273245e-05,
|
| 214 |
+
"loss": 1.9914,
|
| 215 |
+
"step": 950
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.8888888888888888,
|
| 219 |
+
"grad_norm": 13.6875,
|
| 220 |
+
"learning_rate": 3.4040614252052305e-05,
|
| 221 |
+
"loss": 1.9724,
|
| 222 |
+
"step": 1000
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.8888888888888888,
|
| 226 |
+
"eval_loss": 1.9275351762771606,
|
| 227 |
+
"eval_runtime": 48.8023,
|
| 228 |
+
"eval_samples_per_second": 20.491,
|
| 229 |
+
"eval_steps_per_second": 10.245,
|
| 230 |
+
"step": 1000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.9333333333333333,
|
| 234 |
+
"grad_norm": 13.25,
|
| 235 |
+
"learning_rate": 3.2207227887960935e-05,
|
| 236 |
+
"loss": 1.9607,
|
| 237 |
+
"step": 1050
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.9777777777777777,
|
| 241 |
+
"grad_norm": 15.375,
|
| 242 |
+
"learning_rate": 3.033049640731711e-05,
|
| 243 |
+
"loss": 1.9418,
|
| 244 |
+
"step": 1100
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 0.9777777777777777,
|
| 248 |
+
"eval_loss": 1.9146583080291748,
|
| 249 |
+
"eval_runtime": 48.7562,
|
| 250 |
+
"eval_samples_per_second": 20.51,
|
| 251 |
+
"eval_steps_per_second": 10.255,
|
| 252 |
+
"step": 1100
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.0222222222222221,
|
| 256 |
+
"grad_norm": 11.875,
|
| 257 |
+
"learning_rate": 2.8421706694069926e-05,
|
| 258 |
+
"loss": 1.8229,
|
| 259 |
+
"step": 1150
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 1.0666666666666667,
|
| 263 |
+
"grad_norm": 13.625,
|
| 264 |
+
"learning_rate": 2.649233843415149e-05,
|
| 265 |
+
"loss": 1.6831,
|
| 266 |
+
"step": 1200
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.0666666666666667,
|
| 270 |
+
"eval_loss": 1.9151355028152466,
|
| 271 |
+
"eval_runtime": 48.9802,
|
| 272 |
+
"eval_samples_per_second": 20.416,
|
| 273 |
+
"eval_steps_per_second": 10.208,
|
| 274 |
+
"step": 1200
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.1111111111111112,
|
| 278 |
+
"grad_norm": 12.625,
|
| 279 |
+
"learning_rate": 2.4553995075294933e-05,
|
| 280 |
+
"loss": 1.6801,
|
| 281 |
+
"step": 1250
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.1555555555555554,
|
| 285 |
+
"grad_norm": 14.5,
|
| 286 |
+
"learning_rate": 2.2618334042534464e-05,
|
| 287 |
+
"loss": 1.7079,
|
| 288 |
+
"step": 1300
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.1555555555555554,
|
| 292 |
+
"eval_loss": 1.9095146656036377,
|
| 293 |
+
"eval_runtime": 48.6949,
|
| 294 |
+
"eval_samples_per_second": 20.536,
|
| 295 |
+
"eval_steps_per_second": 10.268,
|
| 296 |
+
"step": 1300
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.2,
|
| 300 |
+
"grad_norm": 11.75,
|
| 301 |
+
"learning_rate": 2.0696996629079526e-05,
|
| 302 |
+
"loss": 1.6667,
|
| 303 |
+
"step": 1350
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.2444444444444445,
|
| 307 |
+
"grad_norm": 13.375,
|
| 308 |
+
"learning_rate": 1.880153798420768e-05,
|
| 309 |
+
"loss": 1.6615,
|
| 310 |
+
"step": 1400
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.2444444444444445,
|
| 314 |
+
"eval_loss": 1.9029484987258911,
|
| 315 |
+
"eval_runtime": 48.6447,
|
| 316 |
+
"eval_samples_per_second": 20.557,
|
| 317 |
+
"eval_steps_per_second": 10.279,
|
| 318 |
+
"step": 1400
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.2888888888888888,
|
| 322 |
+
"grad_norm": 15.3125,
|
| 323 |
+
"learning_rate": 1.6943357619237226e-05,
|
| 324 |
+
"loss": 1.6417,
|
| 325 |
+
"step": 1450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.3333333333333333,
|
| 329 |
+
"grad_norm": 12.75,
|
| 330 |
+
"learning_rate": 1.5133630849524793e-05,
|
| 331 |
+
"loss": 1.6729,
|
| 332 |
+
"step": 1500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.3333333333333333,
|
| 336 |
+
"eval_loss": 1.8957328796386719,
|
| 337 |
+
"eval_runtime": 48.7052,
|
| 338 |
+
"eval_samples_per_second": 20.532,
|
| 339 |
+
"eval_steps_per_second": 10.266,
|
| 340 |
+
"step": 1500
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.3777777777777778,
|
| 344 |
+
"grad_norm": 12.8125,
|
| 345 |
+
"learning_rate": 1.3383241584803884e-05,
|
| 346 |
+
"loss": 1.6703,
|
| 347 |
+
"step": 1550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.4222222222222223,
|
| 351 |
+
"grad_norm": 12.3125,
|
| 352 |
+
"learning_rate": 1.170271687207106e-05,
|
| 353 |
+
"loss": 1.6666,
|
| 354 |
+
"step": 1600
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.4222222222222223,
|
| 358 |
+
"eval_loss": 1.89302659034729,
|
| 359 |
+
"eval_runtime": 48.6027,
|
| 360 |
+
"eval_samples_per_second": 20.575,
|
| 361 |
+
"eval_steps_per_second": 10.287,
|
| 362 |
+
"step": 1600
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.4666666666666668,
|
| 366 |
+
"grad_norm": 15.5,
|
| 367 |
+
"learning_rate": 1.010216358468665e-05,
|
| 368 |
+
"loss": 1.6447,
|
| 369 |
+
"step": 1650
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.511111111111111,
|
| 373 |
+
"grad_norm": 12.625,
|
| 374 |
+
"learning_rate": 8.591207638449154e-06,
|
| 375 |
+
"loss": 1.6755,
|
| 376 |
+
"step": 1700
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.511111111111111,
|
| 380 |
+
"eval_loss": 1.891126275062561,
|
| 381 |
+
"eval_runtime": 48.5903,
|
| 382 |
+
"eval_samples_per_second": 20.58,
|
| 383 |
+
"eval_steps_per_second": 10.29,
|
| 384 |
+
"step": 1700
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 1.5555555555555556,
|
| 388 |
+
"grad_norm": 14.0625,
|
| 389 |
+
"learning_rate": 7.178936100204994e-06,
|
| 390 |
+
"loss": 1.6474,
|
| 391 |
+
"step": 1750
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 1.6,
|
| 395 |
+
"grad_norm": 12.6875,
|
| 396 |
+
"learning_rate": 5.873842537159274e-06,
|
| 397 |
+
"loss": 1.66,
|
| 398 |
+
"step": 1800
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 1.6,
|
| 402 |
+
"eval_loss": 1.8891392946243286,
|
| 403 |
+
"eval_runtime": 48.6956,
|
| 404 |
+
"eval_samples_per_second": 20.536,
|
| 405 |
+
"eval_steps_per_second": 10.268,
|
| 406 |
+
"step": 1800
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 1.6444444444444444,
|
| 410 |
+
"grad_norm": 14.0625,
|
| 411 |
+
"learning_rate": 4.683775935563222e-06,
|
| 412 |
+
"loss": 1.682,
|
| 413 |
+
"step": 1850
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.6888888888888889,
|
| 417 |
+
"grad_norm": 14.6875,
|
| 418 |
+
"learning_rate": 3.6158934959873353e-06,
|
| 419 |
+
"loss": 1.6299,
|
| 420 |
+
"step": 1900
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 1.6888888888888889,
|
| 424 |
+
"eval_loss": 1.8877633810043335,
|
| 425 |
+
"eval_runtime": 48.685,
|
| 426 |
+
"eval_samples_per_second": 20.54,
|
| 427 |
+
"eval_steps_per_second": 10.27,
|
| 428 |
+
"step": 1900
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 1.7333333333333334,
|
| 432 |
+
"grad_norm": 11.875,
|
| 433 |
+
"learning_rate": 2.6766175890749786e-06,
|
| 434 |
+
"loss": 1.6601,
|
| 435 |
+
"step": 1950
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 1.7777777777777777,
|
| 439 |
+
"grad_norm": 13.625,
|
| 440 |
+
"learning_rate": 1.8715971306496745e-06,
|
| 441 |
+
"loss": 1.6286,
|
| 442 |
+
"step": 2000
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 1.7777777777777777,
|
| 446 |
+
"eval_loss": 1.8876391649246216,
|
| 447 |
+
"eval_runtime": 48.4452,
|
| 448 |
+
"eval_samples_per_second": 20.642,
|
| 449 |
+
"eval_steps_per_second": 10.321,
|
| 450 |
+
"step": 2000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.8222222222222222,
|
| 454 |
+
"grad_norm": 11.625,
|
| 455 |
+
"learning_rate": 1.2056736084706589e-06,
|
| 456 |
+
"loss": 1.6788,
|
| 457 |
+
"step": 2050
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.8666666666666667,
|
| 461 |
+
"grad_norm": 13.125,
|
| 462 |
+
"learning_rate": 6.828519649558191e-07,
|
| 463 |
+
"loss": 1.6811,
|
| 464 |
+
"step": 2100
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.8666666666666667,
|
| 468 |
+
"eval_loss": 1.8877239227294922,
|
| 469 |
+
"eval_runtime": 48.6643,
|
| 470 |
+
"eval_samples_per_second": 20.549,
|
| 471 |
+
"eval_steps_per_second": 10.274,
|
| 472 |
+
"step": 2100
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.911111111111111,
|
| 476 |
+
"grad_norm": 12.1875,
|
| 477 |
+
"learning_rate": 3.062765109867499e-07,
|
| 478 |
+
"loss": 1.6562,
|
| 479 |
+
"step": 2150
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.9555555555555557,
|
| 483 |
+
"grad_norm": 13.4375,
|
| 484 |
+
"learning_rate": 7.821201565316182e-08,
|
| 485 |
+
"loss": 1.6527,
|
| 486 |
+
"step": 2200
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.9555555555555557,
|
| 490 |
+
"eval_loss": 1.8878294229507446,
|
| 491 |
+
"eval_runtime": 48.65,
|
| 492 |
+
"eval_samples_per_second": 20.555,
|
| 493 |
+
"eval_steps_per_second": 10.277,
|
| 494 |
+
"step": 2200
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.0,
|
| 498 |
+
"grad_norm": 13.75,
|
| 499 |
+
"learning_rate": 3.008566505646737e-11,
|
| 500 |
+
"loss": 1.6683,
|
| 501 |
+
"step": 2250
|
| 502 |
+
}
|
| 503 |
+
],
|
| 504 |
+
"logging_steps": 50,
|
| 505 |
+
"max_steps": 2250,
|
| 506 |
+
"num_input_tokens_seen": 0,
|
| 507 |
+
"num_train_epochs": 2,
|
| 508 |
+
"save_steps": 200,
|
| 509 |
+
"stateful_callbacks": {
|
| 510 |
+
"TrainerControl": {
|
| 511 |
+
"args": {
|
| 512 |
+
"should_epoch_stop": false,
|
| 513 |
+
"should_evaluate": false,
|
| 514 |
+
"should_log": false,
|
| 515 |
+
"should_save": true,
|
| 516 |
+
"should_training_stop": true
|
| 517 |
+
},
|
| 518 |
+
"attributes": {}
|
| 519 |
+
}
|
| 520 |
+
},
|
| 521 |
+
"total_flos": 2773813690368000.0,
|
| 522 |
+
"train_batch_size": 2,
|
| 523 |
+
"trial_name": null,
|
| 524 |
+
"trial_params": null
|
| 525 |
+
}
|
checkpoint-2250/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51ff9f61bcdaae5ac53531ca1cc31aab96bc0e3a0c0ccd88418c58ca3f44c82d
|
| 3 |
+
size 5304
|
config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_sliding_window_pattern": 6,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"Gemma3ForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"attn_logit_softcapping": null,
|
| 9 |
+
"bos_token_id": 2,
|
| 10 |
+
"cache_implementation": "hybrid",
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"final_logit_softcapping": null,
|
| 13 |
+
"head_dim": 256,
|
| 14 |
+
"hidden_activation": "gelu_pytorch_tanh",
|
| 15 |
+
"hidden_size": 640,
|
| 16 |
+
"initializer_range": 0.02,
|
| 17 |
+
"intermediate_size": 2048,
|
| 18 |
+
"layer_types": [
|
| 19 |
+
"sliding_attention",
|
| 20 |
+
"sliding_attention",
|
| 21 |
+
"sliding_attention",
|
| 22 |
+
"sliding_attention",
|
| 23 |
+
"sliding_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"sliding_attention",
|
| 26 |
+
"sliding_attention",
|
| 27 |
+
"sliding_attention",
|
| 28 |
+
"sliding_attention",
|
| 29 |
+
"sliding_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"sliding_attention",
|
| 32 |
+
"sliding_attention",
|
| 33 |
+
"sliding_attention",
|
| 34 |
+
"sliding_attention",
|
| 35 |
+
"sliding_attention",
|
| 36 |
+
"full_attention"
|
| 37 |
+
],
|
| 38 |
+
"max_position_embeddings": 32768,
|
| 39 |
+
"model_type": "gemma3_text",
|
| 40 |
+
"num_attention_heads": 4,
|
| 41 |
+
"num_hidden_layers": 18,
|
| 42 |
+
"num_key_value_heads": 1,
|
| 43 |
+
"pad_token_id": 0,
|
| 44 |
+
"query_pre_attn_scalar": 256,
|
| 45 |
+
"rms_norm_eps": 1e-06,
|
| 46 |
+
"rope_local_base_freq": 10000.0,
|
| 47 |
+
"rope_scaling": null,
|
| 48 |
+
"rope_theta": 1000000.0,
|
| 49 |
+
"sliding_window": 512,
|
| 50 |
+
"sliding_window_pattern": 6,
|
| 51 |
+
"torch_dtype": "bfloat16",
|
| 52 |
+
"transformers_version": "4.52.4",
|
| 53 |
+
"use_bidirectional_attention": false,
|
| 54 |
+
"use_cache": true,
|
| 55 |
+
"vocab_size": 262144
|
| 56 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cache_implementation": "hybrid",
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
1,
|
| 6 |
+
106
|
| 7 |
+
],
|
| 8 |
+
"top_k": 64,
|
| 9 |
+
"top_p": 0.95,
|
| 10 |
+
"transformers_version": "4.52.4"
|
| 11 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7204dfb6f3a944a032ca5ed20d71c923a4fed2c3b11dfabc2c1ff9f2fabe8af8
|
| 3 |
+
size 536223056
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"boi_token": "<start_of_image>",
|
| 3 |
+
"bos_token": {
|
| 4 |
+
"content": "<bos>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false
|
| 9 |
+
},
|
| 10 |
+
"eoi_token": "<end_of_image>",
|
| 11 |
+
"eos_token": {
|
| 12 |
+
"content": "<eos>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false
|
| 17 |
+
},
|
| 18 |
+
"image_token": "<image_soft_token>",
|
| 19 |
+
"pad_token": {
|
| 20 |
+
"content": "<pad>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false
|
| 25 |
+
},
|
| 26 |
+
"unk_token": {
|
| 27 |
+
"content": "<unk>",
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"normalized": false,
|
| 30 |
+
"rstrip": false,
|
| 31 |
+
"single_word": false
|
| 32 |
+
}
|
| 33 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e570d6288ff0afcea981a80492eddfa3e2239a79de89e5074cbb74b548fa5e2b
|
| 3 |
+
size 33384833
|