Model save

README.md

@@ -0,0 +1,60 @@
+---
+license: apache-2.0
+base_model: alignment-handbook/zephyr-7b-dpo-full
+tags:
+- generated_from_trainer
+model-index:
+- name: zephyr-7b-sft-lora
+  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. -->
+
+# zephyr-7b-sft-lora
+
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-dpo-full](https://huggingface.co/alignment-handbook/zephyr-7b-dpo-full) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.7803
+
+## 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: 4
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- gradient_accumulation_steps: 128
+- total_train_batch_size: 512
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- num_epochs: 1
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 2.0318        | 0.32  | 10   | 1.7803          |
+
+
+### Framework versions
+
+- Transformers 4.35.0
+- Pytorch 2.1.0
+- Datasets 2.14.6
+- Tokenizers 0.14.1

adapter_config.json

@@ -8,21 +8,18 @@
   "init_lora_weights": true,
   "layers_pattern": null,
   "layers_to_transform": null,
-  "loftq_config": {},
   "lora_alpha": 16,
-  "lora_dropout": 0.05,
-  "megatron_config": null,
-  "megatron_core": "megatron.core",
+  "lora_dropout": 0.1,
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 16,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
+    "q_proj",
     "k_proj",
     "o_proj",
-    "q_proj"
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fb72397b68e1c9b79a75ee2f0d37aef3012f4e3e9f393e220859b30e1dccde24
-size 27297544
+oid sha256:7632c5ef709adc03b83f1f32f43d9dd75f3294bca163ead95e2d1731d9130310
+size 218138576

all_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 0.32,
+    "eval_loss": 1.7802537679672241,
+    "eval_runtime": 153.411,
+    "eval_samples": 1769,
+    "eval_samples_per_second": 11.531,
+    "eval_steps_per_second": 1.447,
+    "train_loss": 2.409137284755707,
+    "train_runtime": 4600.7142,
+    "train_samples": 15899,
+    "train_samples_per_second": 3.456,
+    "train_steps_per_second": 0.007
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.32,
+    "eval_loss": 1.7802537679672241,
+    "eval_runtime": 153.411,
+    "eval_samples": 1769,
+    "eval_samples_per_second": 11.531,
+    "eval_steps_per_second": 1.447
+}

tokenizer_config.json

@@ -1,6 +1,4 @@
 {
-  "add_bos_token": true,
-  "add_eos_token": false,
   "added_tokens_decoder": {
     "0": {
       "content": "<unk>",
@@ -29,7 +27,7 @@
   },
   "additional_special_tokens": [],
   "bos_token": "<s>",
-  "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}",
+  "chat_template": "{% for message in messages %}\n{% if message['role'] == 'user' %}\n{{ '<|user|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'system' %}\n{{ '<|system|>\n' + message['content'] + eos_token }}\n{% elif message['role'] == 'assistant' %}\n{{ '<|assistant|>\n'  + message['content'] + eos_token }}\n{% endif %}\n{% if loop.last and add_generation_prompt %}\n{{ '<|assistant|>' }}\n{% endif %}\n{% endfor %}",
   "clean_up_tokenization_spaces": false,
   "eos_token": "</s>",
   "legacy": true,

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 0.32,
+    "train_loss": 2.409137284755707,
+    "train_runtime": 4600.7142,
+    "train_samples": 15899,
+    "train_samples_per_second": 3.456,
+    "train_steps_per_second": 0.007
+}

trainer_state.json

@@ -0,0 +1,54 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3220125786163522,
+  "eval_steps": 500,
+  "global_step": 10,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.994869323391895e-05,
+      "loss": 2.5826,
+      "step": 1
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.8743466161445823e-05,
+      "loss": 2.4436,
+      "step": 5
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5289640103269626e-05,
+      "loss": 2.0318,
+      "step": 10
+    },
+    {
+      "epoch": 0.32,
+      "eval_loss": 1.780266523361206,
+      "eval_runtime": 154.3285,
+      "eval_samples_per_second": 11.463,
+      "eval_steps_per_second": 1.438,
+      "step": 10
+    },
+    {
+      "epoch": 0.32,
+      "step": 10,
+      "total_flos": 4.903271030325248e+17,
+      "train_loss": 2.409137284755707,
+      "train_runtime": 4600.7142,
+      "train_samples_per_second": 3.456,
+      "train_steps_per_second": 0.007
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 31,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "total_flos": 4.903271030325248e+17,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d247e6edadd8eec25fafda911325a153e7ec9bfb084946dbb826f80e3be6ebd
-size 4792
+oid sha256:4f87f906fe3f0905c5f58981519e7e15b12915c522bece020f0cea4c22413af6
+size 4664