Upload 20 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,60 @@
----
-license: mit
----
+---
+library_name: peft
+license: other
+base_model: google/gemma-3-12b-it
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+model-index:
+- name: train_2025-06-07-17-55-57
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# train_2025-06-07-17-55-57
+
+This model is a fine-tuned version of [google/gemma-3-12b-it](https://huggingface.co/google/gemma-3-12b-it) on the top_10_training_dataset and the top_10_validation_dataset datasets.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 2e-05
+- train_batch_size: 5
+- eval_batch_size: 8
+- seed: 42
+- gradient_accumulation_steps: 5
+- total_train_batch_size: 25
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine_with_restarts
+- lr_scheduler_warmup_steps: 150
+- num_epochs: 4.0
+- label_smoothing_factor: 0.05
+
+### Training results
+
+
+
+### Framework versions
+
+- PEFT 0.15.1
+- Transformers 4.51.3
+- Pytorch 2.6.0+cu124
+- Datasets 3.5.0
+- Tokenizers 0.21.1

adapter_config.json

@@ -0,0 +1,180 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "google/gemma-3-12b-it",
+  "bias": "none",
+  "corda_config": null,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 64,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 32,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "model.layers.16.self_attn.k_proj",
+    "model.layers.14.self_attn.q_proj",
+    "model.layers.22.self_attn.v_proj",
+    "38.self_attn.v_proj",
+    "40.self_attn.v_proj",
+    "model.layers.24.self_attn.v_proj",
+    "37.self_attn.v_proj",
+    "model.layers.26.self_attn.v_proj",
+    "34.self_attn.k_proj",
+    "45.self_attn.v_proj",
+    "29.self_attn.v_proj",
+    "model.layers.4.self_attn.q_proj",
+    "model.layers.3.self_attn.k_proj",
+    "34.self_attn.v_proj",
+    "model.layers.15.self_attn.q_proj",
+    "model.layers.5.self_attn.v_proj",
+    "model.layers.19.self_attn.v_proj",
+    "45.self_attn.q_proj",
+    "model.layers.20.self_attn.k_proj",
+    "model.layers.5.self_attn.k_proj",
+    "model.layers.17.self_attn.k_proj",
+    "model.layers.26.self_attn.k_proj",
+    "model.layers.8.self_attn.v_proj",
+    "model.layers.7.self_attn.v_proj",
+    "model.layers.14.self_attn.k_proj",
+    "44.self_attn.k_proj",
+    "model.layers.18.self_attn.v_proj",
+    "31.self_attn.k_proj",
+    "44.self_attn.q_proj",
+    "32.self_attn.v_proj",
+    "27.self_attn.q_proj",
+    "47.self_attn.k_proj",
+    "model.layers.13.self_attn.v_proj",
+    "model.layers.17.self_attn.v_proj",
+    "model.layers.18.self_attn.q_proj",
+    "gate_proj",
+    "model.layers.15.self_attn.v_proj",
+    "model.layers.10.self_attn.k_proj",
+    "model.layers.19.self_attn.k_proj",
+    "up_proj",
+    "model.layers.12.self_attn.v_proj",
+    "47.self_attn.q_proj",
+    "model.layers.8.self_attn.q_proj",
+    "model.layers.23.self_attn.v_proj",
+    "model.layers.3.self_attn.v_proj",
+    "28.self_attn.v_proj",
+    "29.self_attn.k_proj",
+    "45.self_attn.k_proj",
+    "model.layers.25.self_attn.q_proj",
+    "32.self_attn.k_proj",
+    "42.self_attn.v_proj",
+    "model.layers.6.self_attn.q_proj",
+    "30.self_attn.k_proj",
+    "model.layers.14.self_attn.v_proj",
+    "model.layers.23.self_attn.q_proj",
+    "model.layers.1.self_attn.k_proj",
+    "model.layers.16.self_attn.v_proj",
+    "model.layers.0.self_attn.v_proj",
+    "43.self_attn.k_proj",
+    "model.layers.20.self_attn.q_proj",
+    "33.self_attn.q_proj",
+    "36.self_attn.q_proj",
+    "47.self_attn.v_proj",
+    "model.layers.23.self_attn.k_proj",
+    "model.layers.24.self_attn.q_proj",
+    "31.self_attn.q_proj",
+    "43.self_attn.q_proj",
+    "41.self_attn.v_proj",
+    "46.self_attn.q_proj",
+    "35.self_attn.v_proj",
+    "model.layers.12.self_attn.q_proj",
+    "44.self_attn.v_proj",
+    "29.self_attn.q_proj",
+    "model.layers.11.self_attn.k_proj",
+    "37.self_attn.k_proj",
+    "model.layers.3.self_attn.q_proj",
+    "model.layers.13.self_attn.q_proj",
+    "model.layers.7.self_attn.k_proj",
+    "38.self_attn.q_proj",
+    "model.layers.5.self_attn.q_proj",
+    "o_proj",
+    "model.layers.20.self_attn.v_proj",
+    "model.layers.16.self_attn.q_proj",
+    "model.layers.22.self_attn.q_proj",
+    "model.layers.12.self_attn.k_proj",
+    "down_proj",
+    "model.layers.13.self_attn.k_proj",
+    "41.self_attn.k_proj",
+    "model.layers.9.self_attn.v_proj",
+    "model.layers.2.self_attn.k_proj",
+    "31.self_attn.v_proj",
+    "model.layers.25.self_attn.v_proj",
+    "model.layers.22.self_attn.k_proj",
+    "34.self_attn.q_proj",
+    "28.self_attn.k_proj",
+    "model.layers.0.self_attn.q_proj",
+    "27.self_attn.v_proj",
+    "model.layers.21.self_attn.k_proj",
+    "model.layers.10.self_attn.v_proj",
+    "35.self_attn.k_proj",
+    "model.layers.15.self_attn.k_proj",
+    "model.layers.21.self_attn.v_proj",
+    "model.layers.1.self_attn.q_proj",
+    "model.layers.4.self_attn.v_proj",
+    "model.layers.2.self_attn.q_proj",
+    "model.layers.26.self_attn.q_proj",
+    "model.layers.9.self_attn.k_proj",
+    "model.layers.9.self_attn.q_proj",
+    "model.layers.25.self_attn.k_proj",
+    "41.self_attn.q_proj",
+    "30.self_attn.v_proj",
+    "33.self_attn.v_proj",
+    "model.layers.11.self_attn.q_proj",
+    "32.self_attn.q_proj",
+    "model.layers.18.self_attn.k_proj",
+    "46.self_attn.v_proj",
+    "40.self_attn.k_proj",
+    "model.layers.8.self_attn.k_proj",
+    "model.layers.7.self_attn.q_proj",
+    "model.layers.4.self_attn.k_proj",
+    "39.self_attn.k_proj",
+    "36.self_attn.k_proj",
+    "model.layers.0.self_attn.k_proj",
+    "36.self_attn.v_proj",
+    "43.self_attn.v_proj",
+    "model.layers.11.self_attn.v_proj",
+    "model.layers.1.self_attn.v_proj",
+    "42.self_attn.k_proj",
+    "35.self_attn.q_proj",
+    "46.self_attn.k_proj",
+    "model.layers.2.self_attn.v_proj",
+    "model.layers.6.self_attn.k_proj",
+    "model.layers.10.self_attn.q_proj",
+    "model.layers.6.self_attn.v_proj",
+    "39.self_attn.v_proj",
+    "38.self_attn.k_proj",
+    "model.layers.24.self_attn.k_proj",
+    "model.layers.21.self_attn.q_proj",
+    "model.layers.17.self_attn.q_proj",
+    "27.self_attn.k_proj",
+    "model.layers.19.self_attn.q_proj",
+    "28.self_attn.q_proj",
+    "33.self_attn.k_proj",
+    "39.self_attn.q_proj",
+    "40.self_attn.q_proj",
+    "37.self_attn.q_proj",
+    "42.self_attn.q_proj",
+    "30.self_attn.q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75492779c6ec99442914be3d7f5e34fc681e7784438378946c127087605660be
+size 523864568

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<image_soft_token>": 262144
+}

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.99728014505893,
+    "num_input_tokens_seen": 11278520,
+    "total_flos": 7.651349314204147e+17,
+    "train_loss": 1.351207665421746,
+    "train_runtime": 8552.2085,
+    "train_samples_per_second": 2.578,
+    "train_steps_per_second": 0.103
+}

chat_template.json

@@ -0,0 +1,3 @@
+{
+  "chat_template": "{{ bos_token }}\n{%- if messages[0]['role'] == 'system' -%}\n    {%- if messages[0]['content'] is string -%}\n        {%- set first_user_prefix = messages[0]['content'] + '\n\n' -%}\n    {%- else -%}\n        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '\n\n' -%}\n    {%- endif -%}\n    {%- set loop_messages = messages[1:] -%}\n{%- else -%}\n    {%- set first_user_prefix = \"\" -%}\n    {%- set loop_messages = messages -%}\n{%- endif -%}\n{%- for message in loop_messages -%}\n    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) -%}\n        {{ raise_exception(\"Conversation roles must alternate user/assistant/user/assistant/...\") }}\n    {%- endif -%}\n    {%- if (message['role'] == 'assistant') -%}\n        {%- set role = \"model\" -%}\n    {%- else -%}\n        {%- set role = message['role'] -%}\n    {%- endif -%}\n    {{ '<start_of_turn>' + role + '\n' + (first_user_prefix if loop.first else \"\") }}\n    {%- if message['content'] is string -%}\n        {{ message['content'] | trim }}\n    {%- elif message['content'] is iterable -%}\n        {%- for item in message['content'] -%}\n            {%- if item['type'] == 'image' -%}\n                {{ '<start_of_image>' }}\n            {%- elif item['type'] == 'text' -%}\n                {{ item['text'] | trim }}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- else -%}\n        {{ raise_exception(\"Invalid content type\") }}\n    {%- endif -%}\n    {{ '<end_of_turn>\n' }}\n{%- endfor -%}\n{%- if add_generation_prompt -%}\n    {{'<start_of_turn>model\n'}}\n{%- endif -%}\n"
+}

llamaboard_config.yaml

@@ -0,0 +1,79 @@
+top.booster: auto
+top.checkpoint_path: []
+top.finetuning_type: lora
+top.model_name: Gemma-3-12B-Instruct
+top.quantization_bit: none
+top.quantization_method: bnb
+top.rope_scaling: none
+top.template: alpaca
+train.additional_target: ''
+train.apollo_rank: 16
+train.apollo_scale: 32
+train.apollo_target: all
+train.apollo_update_interval: 200
+train.badam_mode: layer
+train.badam_switch_interval: 50
+train.badam_switch_mode: ascending
+train.badam_update_ratio: 0.05
+train.batch_size: 5
+train.compute_type: bf16
+train.create_new_adapter: true
+train.cutoff_len: 512
+train.dataset:
+- top_10_training_dataset
+- top_10_validation_dataset
+train.dataset_dir: data
+train.ds_offload: false
+train.ds_stage: none
+train.extra_args: '{"optim":"adamw_torch","weight_decay":0.01,"label_smoothing_factor":0.05}'
+train.freeze_extra_modules: ''
+train.freeze_trainable_layers: 4
+train.freeze_trainable_modules: all
+train.galore_rank: 16
+train.galore_scale: 2
+train.galore_target: all
+train.galore_update_interval: 200
+train.gradient_accumulation_steps: 5
+train.learning_rate: 2e-5
+train.logging_steps: 25
+train.lora_alpha: 64
+train.lora_dropout: 0.05
+train.lora_rank: 32
+train.lora_target: q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj
+train.loraplus_lr_ratio: 8
+train.lr_scheduler_type: cosine_with_restarts
+train.mask_history: false
+train.max_grad_norm: '0.3'
+train.max_samples: '100000'
+train.neat_packing: false
+train.neftune_alpha: 0
+train.num_train_epochs: '4'
+train.packing: false
+train.ppo_score_norm: false
+train.ppo_whiten_rewards: false
+train.pref_beta: 0.1
+train.pref_ftx: 0
+train.pref_loss: sigmoid
+train.report_to:
+- none
+train.resize_vocab: false
+train.reward_model: []
+train.save_steps: 250
+train.swanlab_api_key: ''
+train.swanlab_link: ''
+train.swanlab_mode: cloud
+train.swanlab_project: llamafactory
+train.swanlab_run_name: ''
+train.swanlab_workspace: ''
+train.train_on_prompt: false
+train.training_stage: Supervised Fine-Tuning
+train.use_apollo: false
+train.use_badam: false
+train.use_dora: false
+train.use_galore: false
+train.use_llama_pro: false
+train.use_pissa: false
+train.use_rslora: false
+train.use_swanlab: false
+train.val_size: 0
+train.warmup_steps: 150

preprocessor_config.json

@@ -0,0 +1,29 @@
+{
+  "do_convert_rgb": null,
+  "do_normalize": true,
+  "do_pan_and_scan": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "Gemma3ImageProcessor",
+  "image_seq_length": 256,
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "pan_and_scan_max_num_crops": null,
+  "pan_and_scan_min_crop_size": null,
+  "pan_and_scan_min_ratio_to_activate": null,
+  "processor_class": "Gemma3Processor",
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 896,
+    "width": 896
+  }
+}

processor_config.json

@@ -0,0 +1,4 @@
+{
+  "image_seq_length": 256,
+  "processor_class": "Gemma3Processor"
+}

special_tokens_map.json

@@ -0,0 +1,33 @@
+{
+  "boi_token": "<start_of_image>",
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eoi_token": "<end_of_image>",
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "image_token": "<image_soft_token>",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4667f2089529e8e7657cfb6d1c19910ae71ff5f28aa7ab2ff2763330affad795
+size 33384568

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1299c11d7cf632ef3b4e11937501358ada021bbdf7c47638d13c0ee982f2e79c
+size 4689074

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.99728014505893,
+    "num_input_tokens_seen": 11278520,
+    "total_flos": 7.651349314204147e+17,
+    "train_loss": 1.351207665421746,
+    "train_runtime": 8552.2085,
+    "train_samples_per_second": 2.578,
+    "train_steps_per_second": 0.103
+}

trainer_log.jsonl

@@ -0,0 +1,36 @@
+{"current_steps": 25, "total_steps": 880, "loss": 6.768, "lr": 3.2000000000000003e-06, "epoch": 0.11332728921124206, "percentage": 2.84, "elapsed_time": "0:04:01", "remaining_time": "2:17:49", "throughput": 1323.46, "total_tokens": 320000}
+{"current_steps": 50, "total_steps": 880, "loss": 2.2714, "lr": 6.533333333333334e-06, "epoch": 0.22665457842248413, "percentage": 5.68, "elapsed_time": "0:08:03", "remaining_time": "2:13:39", "throughput": 1324.71, "total_tokens": 640000}
+{"current_steps": 75, "total_steps": 880, "loss": 1.6663, "lr": 9.866666666666668e-06, "epoch": 0.3399818676337262, "percentage": 8.52, "elapsed_time": "0:12:04", "remaining_time": "2:09:36", "throughput": 1324.97, "total_tokens": 960000}
+{"current_steps": 100, "total_steps": 880, "loss": 1.5351, "lr": 1.3200000000000002e-05, "epoch": 0.45330915684496825, "percentage": 11.36, "elapsed_time": "0:16:05", "remaining_time": "2:05:33", "throughput": 1325.22, "total_tokens": 1280000}
+{"current_steps": 125, "total_steps": 880, "loss": 1.4698, "lr": 1.6533333333333333e-05, "epoch": 0.5666364460562103, "percentage": 14.2, "elapsed_time": "0:20:07", "remaining_time": "2:01:31", "throughput": 1325.37, "total_tokens": 1600000}
+{"current_steps": 150, "total_steps": 880, "loss": 1.4164, "lr": 1.9866666666666667e-05, "epoch": 0.6799637352674524, "percentage": 17.05, "elapsed_time": "0:24:08", "remaining_time": "1:57:29", "throughput": 1325.44, "total_tokens": 1920000}
+{"current_steps": 175, "total_steps": 880, "loss": 1.3952, "lr": 1.994670819911521e-05, "epoch": 0.7932910244786945, "percentage": 19.89, "elapsed_time": "0:28:09", "remaining_time": "1:53:27", "throughput": 1325.53, "total_tokens": 2239840}
+{"current_steps": 200, "total_steps": 880, "loss": 1.3792, "lr": 1.977848341505657e-05, "epoch": 0.9066183136899365, "percentage": 22.73, "elapsed_time": "0:32:11", "remaining_time": "1:49:25", "throughput": 1325.63, "total_tokens": 2559800}
+{"current_steps": 225, "total_steps": 880, "loss": 1.3537, "lr": 1.949717842791432e-05, "epoch": 1.0226654578422485, "percentage": 25.57, "elapsed_time": "0:36:16", "remaining_time": "1:45:34", "throughput": 1325.71, "total_tokens": 2884920}
+{"current_steps": 250, "total_steps": 880, "loss": 1.2956, "lr": 1.9106046300942165e-05, "epoch": 1.1359927470534905, "percentage": 28.41, "elapsed_time": "0:40:17", "remaining_time": "1:41:31", "throughput": 1325.77, "total_tokens": 3204920}
+{"current_steps": 275, "total_steps": 880, "loss": 1.2642, "lr": 1.8609610158889943e-05, "epoch": 1.2493200362647325, "percentage": 31.25, "elapsed_time": "0:44:31", "remaining_time": "1:37:57", "throughput": 1319.47, "total_tokens": 3524920}
+{"current_steps": 300, "total_steps": 880, "loss": 1.2527, "lr": 1.8013610881746767e-05, "epoch": 1.3626473254759746, "percentage": 34.09, "elapsed_time": "0:48:32", "remaining_time": "1:33:51", "throughput": 1320.02, "total_tokens": 3844920}
+{"current_steps": 325, "total_steps": 880, "loss": 1.2602, "lr": 1.732494071613579e-05, "epoch": 1.4759746146872166, "percentage": 36.93, "elapsed_time": "0:52:34", "remaining_time": "1:29:46", "throughput": 1320.49, "total_tokens": 4164920}
+{"current_steps": 350, "total_steps": 880, "loss": 1.2551, "lr": 1.6551563572090855e-05, "epoch": 1.5893019038984586, "percentage": 39.77, "elapsed_time": "0:56:35", "remaining_time": "1:25:41", "throughput": 1320.88, "total_tokens": 4484840}
+{"current_steps": 375, "total_steps": 880, "loss": 1.2325, "lr": 1.5702422926917872e-05, "epoch": 1.7026291931097008, "percentage": 42.61, "elapsed_time": "1:00:36", "remaining_time": "1:21:37", "throughput": 1321.22, "total_tokens": 4804840}
+{"current_steps": 400, "total_steps": 880, "loss": 1.2436, "lr": 1.4787338401157888e-05, "epoch": 1.8159564823209429, "percentage": 45.45, "elapsed_time": "1:04:37", "remaining_time": "1:17:33", "throughput": 1321.53, "total_tokens": 5124840}
+{"current_steps": 425, "total_steps": 880, "loss": 1.2319, "lr": 1.3816892202666591e-05, "epoch": 1.929283771532185, "percentage": 48.3, "elapsed_time": "1:08:39", "remaining_time": "1:13:30", "throughput": 1321.8, "total_tokens": 5444840}
+{"current_steps": 450, "total_steps": 880, "loss": 1.2132, "lr": 1.2802306751992163e-05, "epoch": 2.045330915684497, "percentage": 51.14, "elapsed_time": "1:12:44", "remaining_time": "1:09:30", "throughput": 1322.03, "total_tokens": 5769800}
+{"current_steps": 475, "total_steps": 880, "loss": 1.0794, "lr": 1.1755314904214284e-05, "epoch": 2.158658204895739, "percentage": 53.98, "elapsed_time": "1:16:45", "remaining_time": "1:05:26", "throughput": 1322.25, "total_tokens": 6089800}
+{"current_steps": 500, "total_steps": 880, "loss": 1.0779, "lr": 1.06880242680232e-05, "epoch": 2.271985494106981, "percentage": 56.82, "elapsed_time": "1:20:46", "remaining_time": "1:01:23", "throughput": 1322.44, "total_tokens": 6409680}
+{"current_steps": 525, "total_steps": 880, "loss": 1.0722, "lr": 9.612777191078257e-06, "epoch": 2.385312783318223, "percentage": 59.66, "elapsed_time": "1:24:58", "remaining_time": "0:57:27", "throughput": 1320.03, "total_tokens": 6729680}
+{"current_steps": 550, "total_steps": 880, "loss": 1.0663, "lr": 8.542008030801254e-06, "epoch": 2.498640072529465, "percentage": 62.5, "elapsed_time": "1:28:59", "remaining_time": "0:53:23", "throughput": 1320.33, "total_tokens": 7049680}
+{"current_steps": 575, "total_steps": 880, "loss": 1.056, "lr": 7.4880993611518095e-06, "epoch": 2.611967361740707, "percentage": 65.34, "elapsed_time": "1:33:00", "remaining_time": "0:49:20", "throughput": 1320.6, "total_tokens": 7369680}
+{"current_steps": 600, "total_steps": 880, "loss": 1.0578, "lr": 6.463238778236287e-06, "epoch": 2.725294650951949, "percentage": 68.18, "elapsed_time": "1:37:02", "remaining_time": "0:45:17", "throughput": 1320.75, "total_tokens": 7689680}
+{"current_steps": 625, "total_steps": 880, "loss": 1.0531, "lr": 5.479277960676959e-06, "epoch": 2.838621940163191, "percentage": 71.02, "elapsed_time": "1:41:03", "remaining_time": "0:41:13", "throughput": 1320.96, "total_tokens": 8009600}
+{"current_steps": 650, "total_steps": 880, "loss": 1.0523, "lr": 4.547595614593489e-06, "epoch": 2.951949229374433, "percentage": 73.86, "elapsed_time": "1:45:04", "remaining_time": "0:37:10", "throughput": 1321.16, "total_tokens": 8329600}
+{"current_steps": 675, "total_steps": 880, "loss": 1.0127, "lr": 3.6789658881265135e-06, "epoch": 3.067996373526745, "percentage": 76.7, "elapsed_time": "1:49:09", "remaining_time": "0:33:09", "throughput": 1321.35, "total_tokens": 8654720}
+{"current_steps": 700, "total_steps": 880, "loss": 0.9245, "lr": 2.883433777182255e-06, "epoch": 3.1813236627379875, "percentage": 79.55, "elapsed_time": "1:53:11", "remaining_time": "0:29:06", "throughput": 1321.51, "total_tokens": 8974520}
+{"current_steps": 725, "total_steps": 880, "loss": 0.918, "lr": 2.170198963229372e-06, "epoch": 3.2946509519492295, "percentage": 82.39, "elapsed_time": "1:57:12", "remaining_time": "0:25:03", "throughput": 1321.69, "total_tokens": 9294520}
+{"current_steps": 750, "total_steps": 880, "loss": 0.9192, "lr": 1.547509426469368e-06, "epoch": 3.4079782411604715, "percentage": 85.23, "elapsed_time": "2:01:13", "remaining_time": "0:21:00", "throughput": 1321.84, "total_tokens": 9614520}
+{"current_steps": 775, "total_steps": 880, "loss": 0.9177, "lr": 1.022566064657663e-06, "epoch": 3.5213055303717136, "percentage": 88.07, "elapsed_time": "2:05:26", "remaining_time": "0:16:59", "throughput": 1319.93, "total_tokens": 9934520}
+{"current_steps": 800, "total_steps": 880, "loss": 0.9152, "lr": 6.01439420581047e-07, "epoch": 3.6346328195829556, "percentage": 90.91, "elapsed_time": "2:09:27", "remaining_time": "0:12:56", "throughput": 1320.12, "total_tokens": 10254520}
+{"current_steps": 825, "total_steps": 880, "loss": 0.917, "lr": 2.889994811704966e-07, "epoch": 3.7479601087941976, "percentage": 93.75, "elapsed_time": "2:13:29", "remaining_time": "0:08:53", "throughput": 1320.3, "total_tokens": 10574520}
+{"current_steps": 850, "total_steps": 880, "loss": 0.912, "lr": 8.885936006545304e-08, "epoch": 3.8612873980054396, "percentage": 96.59, "elapsed_time": "2:17:30", "remaining_time": "0:04:51", "throughput": 1320.48, "total_tokens": 10894520}
+{"current_steps": 875, "total_steps": 880, "loss": 0.9239, "lr": 3.333514894887646e-09, "epoch": 3.9746146872166817, "percentage": 99.43, "elapsed_time": "2:21:31", "remaining_time": "0:00:48", "throughput": 1320.65, "total_tokens": 11214520}
+{"current_steps": 880, "total_steps": 880, "epoch": 3.99728014505893, "percentage": 100.0, "elapsed_time": "2:22:32", "remaining_time": "0:00:00", "throughput": 1318.79, "total_tokens": 11278520}

trainer_state.json

@@ -0,0 +1,324 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.99728014505893,
+  "eval_steps": 500,
+  "global_step": 880,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.11332728921124206,
+      "grad_norm": 12.256142616271973,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 6.768,
+      "num_input_tokens_seen": 320000,
+      "step": 25
+    },
+    {
+      "epoch": 0.22665457842248413,
+      "grad_norm": 14.015169143676758,
+      "learning_rate": 6.533333333333334e-06,
+      "loss": 2.2714,
+      "num_input_tokens_seen": 640000,
+      "step": 50
+    },
+    {
+      "epoch": 0.3399818676337262,
+      "grad_norm": 2.729668140411377,
+      "learning_rate": 9.866666666666668e-06,
+      "loss": 1.6663,
+      "num_input_tokens_seen": 960000,
+      "step": 75
+    },
+    {
+      "epoch": 0.45330915684496825,
+      "grad_norm": 2.5849289894104004,
+      "learning_rate": 1.3200000000000002e-05,
+      "loss": 1.5351,
+      "num_input_tokens_seen": 1280000,
+      "step": 100
+    },
+    {
+      "epoch": 0.5666364460562103,
+      "grad_norm": 2.5873868465423584,
+      "learning_rate": 1.6533333333333333e-05,
+      "loss": 1.4698,
+      "num_input_tokens_seen": 1600000,
+      "step": 125
+    },
+    {
+      "epoch": 0.6799637352674524,
+      "grad_norm": 1.8172498941421509,
+      "learning_rate": 1.9866666666666667e-05,
+      "loss": 1.4164,
+      "num_input_tokens_seen": 1920000,
+      "step": 150
+    },
+    {
+      "epoch": 0.7932910244786945,
+      "grad_norm": 3.378293037414551,
+      "learning_rate": 1.994670819911521e-05,
+      "loss": 1.3952,
+      "num_input_tokens_seen": 2239840,
+      "step": 175
+    },
+    {
+      "epoch": 0.9066183136899365,
+      "grad_norm": 1.4730700254440308,
+      "learning_rate": 1.977848341505657e-05,
+      "loss": 1.3792,
+      "num_input_tokens_seen": 2559800,
+      "step": 200
+    },
+    {
+      "epoch": 1.0226654578422485,
+      "grad_norm": 1.2745240926742554,
+      "learning_rate": 1.949717842791432e-05,
+      "loss": 1.3537,
+      "num_input_tokens_seen": 2884920,
+      "step": 225
+    },
+    {
+      "epoch": 1.1359927470534905,
+      "grad_norm": 1.7677239179611206,
+      "learning_rate": 1.9106046300942165e-05,
+      "loss": 1.2956,
+      "num_input_tokens_seen": 3204920,
+      "step": 250
+    },
+    {
+      "epoch": 1.2493200362647325,
+      "grad_norm": 1.9639768600463867,
+      "learning_rate": 1.8609610158889943e-05,
+      "loss": 1.2642,
+      "num_input_tokens_seen": 3524920,
+      "step": 275
+    },
+    {
+      "epoch": 1.3626473254759746,
+      "grad_norm": 1.738120436668396,
+      "learning_rate": 1.8013610881746767e-05,
+      "loss": 1.2527,
+      "num_input_tokens_seen": 3844920,
+      "step": 300
+    },
+    {
+      "epoch": 1.4759746146872166,
+      "grad_norm": 1.5475760698318481,
+      "learning_rate": 1.732494071613579e-05,
+      "loss": 1.2602,
+      "num_input_tokens_seen": 4164920,
+      "step": 325
+    },
+    {
+      "epoch": 1.5893019038984586,
+      "grad_norm": 1.3410508632659912,
+      "learning_rate": 1.6551563572090855e-05,
+      "loss": 1.2551,
+      "num_input_tokens_seen": 4484840,
+      "step": 350
+    },
+    {
+      "epoch": 1.7026291931097008,
+      "grad_norm": 1.6581236124038696,
+      "learning_rate": 1.5702422926917872e-05,
+      "loss": 1.2325,
+      "num_input_tokens_seen": 4804840,
+      "step": 375
+    },
+    {
+      "epoch": 1.8159564823209429,
+      "grad_norm": 1.8297406435012817,
+      "learning_rate": 1.4787338401157888e-05,
+      "loss": 1.2436,
+      "num_input_tokens_seen": 5124840,
+      "step": 400
+    },
+    {
+      "epoch": 1.929283771532185,
+      "grad_norm": 1.9106981754302979,
+      "learning_rate": 1.3816892202666591e-05,
+      "loss": 1.2319,
+      "num_input_tokens_seen": 5444840,
+      "step": 425
+    },
+    {
+      "epoch": 2.045330915684497,
+      "grad_norm": 1.4120745658874512,
+      "learning_rate": 1.2802306751992163e-05,
+      "loss": 1.2132,
+      "num_input_tokens_seen": 5769800,
+      "step": 450
+    },
+    {
+      "epoch": 2.158658204895739,
+      "grad_norm": 1.541704535484314,
+      "learning_rate": 1.1755314904214284e-05,
+      "loss": 1.0794,
+      "num_input_tokens_seen": 6089800,
+      "step": 475
+    },
+    {
+      "epoch": 2.271985494106981,
+      "grad_norm": 2.239482879638672,
+      "learning_rate": 1.06880242680232e-05,
+      "loss": 1.0779,
+      "num_input_tokens_seen": 6409680,
+      "step": 500
+    },
+    {
+      "epoch": 2.385312783318223,
+      "grad_norm": 2.1483075618743896,
+      "learning_rate": 9.612777191078257e-06,
+      "loss": 1.0722,
+      "num_input_tokens_seen": 6729680,
+      "step": 525
+    },
+    {
+      "epoch": 2.498640072529465,
+      "grad_norm": 1.631958246231079,
+      "learning_rate": 8.542008030801254e-06,
+      "loss": 1.0663,
+      "num_input_tokens_seen": 7049680,
+      "step": 550
+    },
+    {
+      "epoch": 2.611967361740707,
+      "grad_norm": 1.948183298110962,
+      "learning_rate": 7.4880993611518095e-06,
+      "loss": 1.056,
+      "num_input_tokens_seen": 7369680,
+      "step": 575
+    },
+    {
+      "epoch": 2.725294650951949,
+      "grad_norm": 1.8783904314041138,
+      "learning_rate": 6.463238778236287e-06,
+      "loss": 1.0578,
+      "num_input_tokens_seen": 7689680,
+      "step": 600
+    },
+    {
+      "epoch": 2.838621940163191,
+      "grad_norm": 2.3182311058044434,
+      "learning_rate": 5.479277960676959e-06,
+      "loss": 1.0531,
+      "num_input_tokens_seen": 8009600,
+      "step": 625
+    },
+    {
+      "epoch": 2.951949229374433,
+      "grad_norm": 2.483482837677002,
+      "learning_rate": 4.547595614593489e-06,
+      "loss": 1.0523,
+      "num_input_tokens_seen": 8329600,
+      "step": 650
+    },
+    {
+      "epoch": 3.067996373526745,
+      "grad_norm": 1.3066755533218384,
+      "learning_rate": 3.6789658881265135e-06,
+      "loss": 1.0127,
+      "num_input_tokens_seen": 8654720,
+      "step": 675
+    },
+    {
+      "epoch": 3.1813236627379875,
+      "grad_norm": 1.4423686265945435,
+      "learning_rate": 2.883433777182255e-06,
+      "loss": 0.9245,
+      "num_input_tokens_seen": 8974520,
+      "step": 700
+    },
+    {
+      "epoch": 3.2946509519492295,
+      "grad_norm": 2.5524277687072754,
+      "learning_rate": 2.170198963229372e-06,
+      "loss": 0.918,
+      "num_input_tokens_seen": 9294520,
+      "step": 725
+    },
+    {
+      "epoch": 3.4079782411604715,
+      "grad_norm": 1.343592882156372,
+      "learning_rate": 1.547509426469368e-06,
+      "loss": 0.9192,
+      "num_input_tokens_seen": 9614520,
+      "step": 750
+    },
+    {
+      "epoch": 3.5213055303717136,
+      "grad_norm": 1.9703131914138794,
+      "learning_rate": 1.022566064657663e-06,
+      "loss": 0.9177,
+      "num_input_tokens_seen": 9934520,
+      "step": 775
+    },
+    {
+      "epoch": 3.6346328195829556,
+      "grad_norm": 2.417506694793701,
+      "learning_rate": 6.01439420581047e-07,
+      "loss": 0.9152,
+      "num_input_tokens_seen": 10254520,
+      "step": 800
+    },
+    {
+      "epoch": 3.7479601087941976,
+      "grad_norm": 1.23152494430542,
+      "learning_rate": 2.889994811704966e-07,
+      "loss": 0.917,
+      "num_input_tokens_seen": 10574520,
+      "step": 825
+    },
+    {
+      "epoch": 3.8612873980054396,
+      "grad_norm": 1.5819039344787598,
+      "learning_rate": 8.885936006545304e-08,
+      "loss": 0.912,
+      "num_input_tokens_seen": 10894520,
+      "step": 850
+    },
+    {
+      "epoch": 3.9746146872166817,
+      "grad_norm": 1.4086824655532837,
+      "learning_rate": 3.333514894887646e-09,
+      "loss": 0.9239,
+      "num_input_tokens_seen": 11214520,
+      "step": 875
+    },
+    {
+      "epoch": 3.99728014505893,
+      "num_input_tokens_seen": 11278520,
+      "step": 880,
+      "total_flos": 7.651349314204147e+17,
+      "train_loss": 1.351207665421746,
+      "train_runtime": 8552.2085,
+      "train_samples_per_second": 2.578,
+      "train_steps_per_second": 0.103
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 880,
+  "num_input_tokens_seen": 11278520,
+  "num_train_epochs": 4,
+  "save_steps": 250,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.651349314204147e+17,
+  "train_batch_size": 5,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:980d246918b584aa93da5b481907857177d001ae8d29652cd9da329db2bdc8ef
+size 5752

training_args.yaml

@@ -0,0 +1,37 @@
+bf16: true
+create_new_adapter: true
+cutoff_len: 512
+dataset: top_10_training_dataset,top_10_validation_dataset
+dataset_dir: data
+ddp_timeout: 180000000
+do_train: true
+finetuning_type: lora
+flash_attn: auto
+gradient_accumulation_steps: 5
+include_num_input_tokens_seen: true
+label_smoothing_factor: 0.05
+learning_rate: 2.0e-05
+logging_steps: 25
+lora_alpha: 64
+lora_dropout: 0.05
+lora_rank: 32
+lora_target: q_proj,k_proj,v_proj,o_proj,gate_proj,up_proj,down_proj
+loraplus_lr_ratio: 8
+lr_scheduler_type: cosine_with_restarts
+max_grad_norm: 0.3
+max_samples: 100000
+model_name_or_path: google/gemma-3-12b-it
+num_train_epochs: 4.0
+optim: adamw_torch
+output_dir: saves/Gemma-3-12B-Instruct/lora/train_2025-06-07-17-55-57
+packing: false
+per_device_train_batch_size: 5
+plot_loss: true
+preprocessing_num_workers: 16
+report_to: none
+save_steps: 250
+stage: sft
+template: alpaca
+trust_remote_code: true
+warmup_steps: 150
+weight_decay: 0.01