Upload 109 files

qwen72B_models/qwen_72B_d0_iter1_model/README.md

@@ -0,0 +1,68 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B
+model-index:
+- name: iter1_10k_filter_prompt_eval_score_qwen_like_2epoch
+  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. -->
+
+# iter1_10k_filter_prompt_eval_score_qwen_like_2epoch
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B) on the iter1_10k_filter_prompt_eval_score dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6907
+
+## 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: 5e-05
+- train_batch_size: 1
+- eval_batch_size: 1
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 16
+- total_eval_batch_size: 8
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 20
+- num_epochs: 2.0
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.6153        | 0.8696 | 100  | 0.6939          |
+| 0.5431        | 1.7391 | 200  | 0.6907          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.40.1
+- Pytorch 2.3.0+cu121
+- Datasets 2.19.0
+- Tokenizers 0.19.1

qwen72B_models/qwen_72B_d0_iter1_model/adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B",
+  "bias": "none",
+  "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": 16,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "gate_proj",
+    "up_proj",
+    "down_proj",
+    "o_proj",
+    "v_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen72B_models/qwen_72B_d0_iter1_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6653a13d4ea53c6d158c5f3590d43801ffe1c4fc5054d2886f48d8eb7b0edd70
+size 209865840

qwen72B_models/qwen_72B_d0_iter1_model/added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

qwen72B_models/qwen_72B_d0_iter1_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.6907033324241638,
+    "eval_runtime": 30.218,
+    "eval_samples_per_second": 6.751,
+    "eval_steps_per_second": 0.86,
+    "total_flos": 294651875557376.0,
+    "train_loss": 0.6822677301323932,
+    "train_runtime": 1850.6815,
+    "train_samples_per_second": 1.982,
+    "train_steps_per_second": 0.124
+}

qwen72B_models/qwen_72B_d0_iter1_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.6907033324241638,
+    "eval_runtime": 30.218,
+    "eval_samples_per_second": 6.751,
+    "eval_steps_per_second": 0.86
+}

qwen72B_models/qwen_72B_d0_iter1_model/special_tokens_map.json

@@ -0,0 +1,27 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

qwen72B_models/qwen_72B_d0_iter1_model/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|endoftext|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% set system_message = 'You are a helpful assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<|im_start|>system\\n' + system_message + '<|im_end|>\\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|im_start|>user\\n' + content + '<|im_end|>\\n<|im_start|>assistant\\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|endoftext|>' + '\\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

qwen72B_models/qwen_72B_d0_iter1_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 294651875557376.0,
+    "train_loss": 0.6822677301323932,
+    "train_runtime": 1850.6815,
+    "train_samples_per_second": 1.982,
+    "train_steps_per_second": 0.124
+}

qwen72B_models/qwen_72B_d0_iter1_model/trainer_log.jsonl

@@ -0,0 +1,27 @@
+{"current_steps": 10, "total_steps": 230, "loss": 1.2798, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.08695652173913043, "percentage": 4.35, "elapsed_time": "0:01:21", "remaining_time": "0:29:51"}
+{"current_steps": 20, "total_steps": 230, "loss": 0.9826, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.17391304347826086, "percentage": 8.7, "elapsed_time": "0:02:32", "remaining_time": "0:26:36"}
+{"current_steps": 30, "total_steps": 230, "loss": 0.7455, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.972077065562821e-05, "epoch": 0.2608695652173913, "percentage": 13.04, "elapsed_time": "0:03:43", "remaining_time": "0:24:48"}
+{"current_steps": 40, "total_steps": 230, "loss": 0.7768, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.888932014465352e-05, "epoch": 0.34782608695652173, "percentage": 17.39, "elapsed_time": "0:04:53", "remaining_time": "0:23:14"}
+{"current_steps": 50, "total_steps": 230, "loss": 0.651, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.752422169756048e-05, "epoch": 0.43478260869565216, "percentage": 21.74, "elapsed_time": "0:06:03", "remaining_time": "0:21:49"}
+{"current_steps": 60, "total_steps": 230, "loss": 0.6196, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.5655969357899874e-05, "epoch": 0.5217391304347826, "percentage": 26.09, "elapsed_time": "0:07:14", "remaining_time": "0:20:30"}
+{"current_steps": 70, "total_steps": 230, "loss": 0.6283, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.332629679574566e-05, "epoch": 0.6086956521739131, "percentage": 30.43, "elapsed_time": "0:08:24", "remaining_time": "0:19:13"}
+{"current_steps": 80, "total_steps": 230, "loss": 0.6584, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.058724504646834e-05, "epoch": 0.6956521739130435, "percentage": 34.78, "elapsed_time": "0:09:34", "remaining_time": "0:17:57"}
+{"current_steps": 90, "total_steps": 230, "loss": 0.7047, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.7500000000000003e-05, "epoch": 0.782608695652174, "percentage": 39.13, "elapsed_time": "0:10:44", "remaining_time": "0:16:43"}
+{"current_steps": 100, "total_steps": 230, "loss": 0.6153, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.413352560915988e-05, "epoch": 0.8695652173913043, "percentage": 43.48, "elapsed_time": "0:11:54", "remaining_time": "0:15:29"}
+{"current_steps": 100, "total_steps": 230, "loss": null, "eval_loss": 0.6939123868942261, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.8695652173913043, "percentage": 43.48, "elapsed_time": "0:11:54", "remaining_time": "0:15:29"}
+{"current_steps": 110, "total_steps": 230, "loss": 0.6807, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.056302334890786e-05, "epoch": 0.9565217391304348, "percentage": 47.83, "elapsed_time": "0:14:58", "remaining_time": "0:16:19"}
+{"current_steps": 120, "total_steps": 230, "loss": 0.6801, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.686825233966061e-05, "epoch": 1.0434782608695652, "percentage": 52.17, "elapsed_time": "0:16:08", "remaining_time": "0:14:47"}
+{"current_steps": 130, "total_steps": 230, "loss": 0.6434, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.3131747660339394e-05, "epoch": 1.1304347826086956, "percentage": 56.52, "elapsed_time": "0:17:19", "remaining_time": "0:13:19"}
+{"current_steps": 140, "total_steps": 230, "loss": 0.5609, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9436976651092144e-05, "epoch": 1.2173913043478262, "percentage": 60.87, "elapsed_time": "0:18:29", "remaining_time": "0:11:53"}
+{"current_steps": 150, "total_steps": 230, "loss": 0.6344, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.5866474390840125e-05, "epoch": 1.3043478260869565, "percentage": 65.22, "elapsed_time": "0:19:40", "remaining_time": "0:10:29"}
+{"current_steps": 160, "total_steps": 230, "loss": 0.5988, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.2500000000000006e-05, "epoch": 1.391304347826087, "percentage": 69.57, "elapsed_time": "0:20:50", "remaining_time": "0:09:07"}
+{"current_steps": 170, "total_steps": 230, "loss": 0.6445, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.412754953531663e-06, "epoch": 1.4782608695652173, "percentage": 73.91, "elapsed_time": "0:22:00", "remaining_time": "0:07:46"}
+{"current_steps": 180, "total_steps": 230, "loss": 0.6121, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.673703204254347e-06, "epoch": 1.5652173913043477, "percentage": 78.26, "elapsed_time": "0:23:11", "remaining_time": "0:06:26"}
+{"current_steps": 190, "total_steps": 230, "loss": 0.6374, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.344030642100133e-06, "epoch": 1.6521739130434783, "percentage": 82.61, "elapsed_time": "0:24:21", "remaining_time": "0:05:07"}
+{"current_steps": 200, "total_steps": 230, "loss": 0.5431, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.475778302439524e-06, "epoch": 1.7391304347826086, "percentage": 86.96, "elapsed_time": "0:25:30", "remaining_time": "0:03:49"}
+{"current_steps": 200, "total_steps": 230, "loss": null, "eval_loss": 0.6907033324241638, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.7391304347826086, "percentage": 86.96, "elapsed_time": "0:25:30", "remaining_time": "0:03:49"}
+{"current_steps": 210, "total_steps": 230, "loss": 0.601, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.1106798553464804e-06, "epoch": 1.8260869565217392, "percentage": 91.3, "elapsed_time": "0:28:30", "remaining_time": "0:02:42"}
+{"current_steps": 220, "total_steps": 230, "loss": 0.6266, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.7922934437178695e-07, "epoch": 1.9130434782608696, "percentage": 95.65, "elapsed_time": "0:29:39", "remaining_time": "0:01:20"}
+{"current_steps": 230, "total_steps": 230, "loss": 0.5671, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 0.0, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:30:50", "remaining_time": "0:00:00"}
+{"current_steps": 230, "total_steps": 230, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:30:50", "remaining_time": "0:00:00"}
+{"current_steps": 26, "total_steps": 26, "loss": null, "eval_loss": 0.6907033324241638, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:32:39", "remaining_time": "0:00:00"}

qwen72B_models/qwen_72B_d0_iter1_model/trainer_state.json

@@ -0,0 +1,207 @@
+{
+  "best_metric": 0.6907033324241638,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/iter1_10k_filter_prompt_eval_score_qwen_like_2epoch/checkpoint-200",
+  "epoch": 2.0,
+  "eval_steps": 100,
+  "global_step": 230,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.08695652173913043,
+      "grad_norm": 0.7590952864317322,
+      "learning_rate": 2.5e-05,
+      "loss": 1.2798,
+      "step": 10
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "grad_norm": 0.7111290755249571,
+      "learning_rate": 5e-05,
+      "loss": 0.9826,
+      "step": 20
+    },
+    {
+      "epoch": 0.2608695652173913,
+      "grad_norm": 0.6372000562604043,
+      "learning_rate": 4.972077065562821e-05,
+      "loss": 0.7455,
+      "step": 30
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "grad_norm": 0.31441395862223226,
+      "learning_rate": 4.888932014465352e-05,
+      "loss": 0.7768,
+      "step": 40
+    },
+    {
+      "epoch": 0.43478260869565216,
+      "grad_norm": 0.6053951979934512,
+      "learning_rate": 4.752422169756048e-05,
+      "loss": 0.651,
+      "step": 50
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "grad_norm": 0.6427601225513047,
+      "learning_rate": 4.5655969357899874e-05,
+      "loss": 0.6196,
+      "step": 60
+    },
+    {
+      "epoch": 0.6086956521739131,
+      "grad_norm": 0.48016282184342496,
+      "learning_rate": 4.332629679574566e-05,
+      "loss": 0.6283,
+      "step": 70
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "grad_norm": 0.5358203948261187,
+      "learning_rate": 4.058724504646834e-05,
+      "loss": 0.6584,
+      "step": 80
+    },
+    {
+      "epoch": 0.782608695652174,
+      "grad_norm": 0.5267676485714428,
+      "learning_rate": 3.7500000000000003e-05,
+      "loss": 0.7047,
+      "step": 90
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "grad_norm": 1.1981570399779329,
+      "learning_rate": 3.413352560915988e-05,
+      "loss": 0.6153,
+      "step": 100
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "eval_loss": 0.6939123868942261,
+      "eval_runtime": 31.6029,
+      "eval_samples_per_second": 6.455,
+      "eval_steps_per_second": 0.823,
+      "step": 100
+    },
+    {
+      "epoch": 0.9565217391304348,
+      "grad_norm": 0.43612358582278793,
+      "learning_rate": 3.056302334890786e-05,
+      "loss": 0.6807,
+      "step": 110
+    },
+    {
+      "epoch": 1.0434782608695652,
+      "grad_norm": 0.715325390647874,
+      "learning_rate": 2.686825233966061e-05,
+      "loss": 0.6801,
+      "step": 120
+    },
+    {
+      "epoch": 1.1304347826086956,
+      "grad_norm": 0.45417612349207276,
+      "learning_rate": 2.3131747660339394e-05,
+      "loss": 0.6434,
+      "step": 130
+    },
+    {
+      "epoch": 1.2173913043478262,
+      "grad_norm": 0.47181957078682596,
+      "learning_rate": 1.9436976651092144e-05,
+      "loss": 0.5609,
+      "step": 140
+    },
+    {
+      "epoch": 1.3043478260869565,
+      "grad_norm": 0.5021743813880641,
+      "learning_rate": 1.5866474390840125e-05,
+      "loss": 0.6344,
+      "step": 150
+    },
+    {
+      "epoch": 1.391304347826087,
+      "grad_norm": 0.31512508897758795,
+      "learning_rate": 1.2500000000000006e-05,
+      "loss": 0.5988,
+      "step": 160
+    },
+    {
+      "epoch": 1.4782608695652173,
+      "grad_norm": 0.5643879444229492,
+      "learning_rate": 9.412754953531663e-06,
+      "loss": 0.6445,
+      "step": 170
+    },
+    {
+      "epoch": 1.5652173913043477,
+      "grad_norm": 0.4694732084758121,
+      "learning_rate": 6.673703204254347e-06,
+      "loss": 0.6121,
+      "step": 180
+    },
+    {
+      "epoch": 1.6521739130434783,
+      "grad_norm": 0.6643626029082335,
+      "learning_rate": 4.344030642100133e-06,
+      "loss": 0.6374,
+      "step": 190
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "grad_norm": 0.5202263586130002,
+      "learning_rate": 2.475778302439524e-06,
+      "loss": 0.5431,
+      "step": 200
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "eval_loss": 0.6907033324241638,
+      "eval_runtime": 30.0085,
+      "eval_samples_per_second": 6.798,
+      "eval_steps_per_second": 0.866,
+      "step": 200
+    },
+    {
+      "epoch": 1.8260869565217392,
+      "grad_norm": 0.841152000855221,
+      "learning_rate": 1.1106798553464804e-06,
+      "loss": 0.601,
+      "step": 210
+    },
+    {
+      "epoch": 1.9130434782608696,
+      "grad_norm": 0.5051236922199935,
+      "learning_rate": 2.7922934437178695e-07,
+      "loss": 0.6266,
+      "step": 220
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.39770761752834893,
+      "learning_rate": 0.0,
+      "loss": 0.5671,
+      "step": 230
+    },
+    {
+      "epoch": 2.0,
+      "step": 230,
+      "total_flos": 294651875557376.0,
+      "train_loss": 0.6822677301323932,
+      "train_runtime": 1850.6815,
+      "train_samples_per_second": 1.982,
+      "train_steps_per_second": 0.124
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 230,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "total_flos": 294651875557376.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen72B_models/qwen_72B_d0_iter1_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04b0857cde8cf3df437e86e088ed0afebf9079e5bd7fc1df1abbde9d433d317b
+size 6968

qwen72B_models/qwen_72B_d1_iter2_model/README.md

@@ -0,0 +1,69 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B
+model-index:
+- name: on_base_iter2_2epoch_d1
+  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. -->
+
+# on_base_iter2_2epoch_d1
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B) on the on_base_iter2_2epoch_d1 dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.4966
+
+## 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: 5e-05
+- train_batch_size: 1
+- eval_batch_size: 1
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 16
+- total_eval_batch_size: 8
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 20
+- num_epochs: 2.0
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.4598        | 0.5917 | 100  | 0.5030          |
+| 0.4722        | 1.1834 | 200  | 0.4966          |
+| 0.4785        | 1.7751 | 300  | 0.5004          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.40.1
+- Pytorch 2.3.0+cu121
+- Datasets 2.19.0
+- Tokenizers 0.19.1

qwen72B_models/qwen_72B_d1_iter2_model/adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B",
+  "bias": "none",
+  "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": 16,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "q_proj",
+    "k_proj",
+    "v_proj",
+    "down_proj",
+    "up_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen72B_models/qwen_72B_d1_iter2_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2dc3e9cc5429a17504e3158d3e389ceb8cd862ee0caa57cc79d894d61233b57a
+size 209865840

qwen72B_models/qwen_72B_d1_iter2_model/added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

qwen72B_models/qwen_72B_d1_iter2_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.496601402759552,
+    "eval_runtime": 42.8154,
+    "eval_samples_per_second": 7.007,
+    "eval_steps_per_second": 0.888,
+    "total_flos": 312432690462720.0,
+    "train_loss": 0.5284144356405946,
+    "train_runtime": 2754.0946,
+    "train_samples_per_second": 1.961,
+    "train_steps_per_second": 0.123
+}

qwen72B_models/qwen_72B_d1_iter2_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.496601402759552,
+    "eval_runtime": 42.8154,
+    "eval_samples_per_second": 7.007,
+    "eval_steps_per_second": 0.888
+}

qwen72B_models/qwen_72B_d1_iter2_model/special_tokens_map.json

@@ -0,0 +1,27 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

qwen72B_models/qwen_72B_d1_iter2_model/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|endoftext|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% set system_message = 'You are a helpful assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<|im_start|>system\\n' + system_message + '<|im_end|>\\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|im_start|>user\\n' + content + '<|im_end|>\\n<|im_start|>assistant\\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|endoftext|>' + '\\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

qwen72B_models/qwen_72B_d1_iter2_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 312432690462720.0,
+    "train_loss": 0.5284144356405946,
+    "train_runtime": 2754.0946,
+    "train_samples_per_second": 1.961,
+    "train_steps_per_second": 0.123
+}

qwen72B_models/qwen_72B_d1_iter2_model/trainer_log.jsonl

@@ -0,0 +1,38 @@
+{"current_steps": 10, "total_steps": 338, "loss": 1.4138, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.05917159763313609, "percentage": 2.96, "elapsed_time": "0:01:21", "remaining_time": "0:44:31"}
+{"current_steps": 20, "total_steps": 338, "loss": 1.0279, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.11834319526627218, "percentage": 5.92, "elapsed_time": "0:02:31", "remaining_time": "0:40:13"}
+{"current_steps": 30, "total_steps": 338, "loss": 0.6745, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.98781004037916e-05, "epoch": 0.17751479289940827, "percentage": 8.88, "elapsed_time": "0:03:41", "remaining_time": "0:37:59"}
+{"current_steps": 40, "total_steps": 338, "loss": 0.5665, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.951359037609088e-05, "epoch": 0.23668639053254437, "percentage": 11.83, "elapsed_time": "0:04:52", "remaining_time": "0:36:16"}
+{"current_steps": 50, "total_steps": 338, "loss": 0.6138, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.891002460691306e-05, "epoch": 0.2958579881656805, "percentage": 14.79, "elapsed_time": "0:06:01", "remaining_time": "0:34:44"}
+{"current_steps": 60, "total_steps": 338, "loss": 0.5885, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.807328905014201e-05, "epoch": 0.35502958579881655, "percentage": 17.75, "elapsed_time": "0:07:12", "remaining_time": "0:33:25"}
+{"current_steps": 70, "total_steps": 338, "loss": 0.5096, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.7011543523897996e-05, "epoch": 0.41420118343195267, "percentage": 20.71, "elapsed_time": "0:08:23", "remaining_time": "0:32:06"}
+{"current_steps": 80, "total_steps": 338, "loss": 0.5879, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.573514213625505e-05, "epoch": 0.47337278106508873, "percentage": 23.67, "elapsed_time": "0:09:33", "remaining_time": "0:30:51"}
+{"current_steps": 90, "total_steps": 338, "loss": 0.4712, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.425653231231344e-05, "epoch": 0.5325443786982249, "percentage": 26.63, "elapsed_time": "0:10:44", "remaining_time": "0:29:34"}
+{"current_steps": 100, "total_steps": 338, "loss": 0.4598, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.259013340731224e-05, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:11:53", "remaining_time": "0:28:19"}
+{"current_steps": 100, "total_steps": 338, "loss": null, "eval_loss": 0.5029934048652649, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:11:53", "remaining_time": "0:28:19"}
+{"current_steps": 110, "total_steps": 338, "loss": 0.5501, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.075219608954278e-05, "epoch": 0.650887573964497, "percentage": 32.54, "elapsed_time": "0:15:09", "remaining_time": "0:31:24"}
+{"current_steps": 120, "total_steps": 338, "loss": 0.4926, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.876064386435646e-05, "epoch": 0.7100591715976331, "percentage": 35.5, "elapsed_time": "0:16:19", "remaining_time": "0:29:38"}
+{"current_steps": 130, "total_steps": 338, "loss": 0.5475, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.663489828471953e-05, "epoch": 0.7692307692307693, "percentage": 38.46, "elapsed_time": "0:17:29", "remaining_time": "0:27:58"}
+{"current_steps": 140, "total_steps": 338, "loss": 0.4966, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.4395689552855955e-05, "epoch": 0.8284023668639053, "percentage": 41.42, "elapsed_time": "0:18:39", "remaining_time": "0:26:23"}
+{"current_steps": 150, "total_steps": 338, "loss": 0.5472, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.206485435998498e-05, "epoch": 0.8875739644970414, "percentage": 44.38, "elapsed_time": "0:19:49", "remaining_time": "0:24:50"}
+{"current_steps": 160, "total_steps": 338, "loss": 0.5468, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.9665122935613727e-05, "epoch": 0.9467455621301775, "percentage": 47.34, "elapsed_time": "0:21:00", "remaining_time": "0:23:21"}
+{"current_steps": 170, "total_steps": 338, "loss": 0.5135, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.7219897383073373e-05, "epoch": 1.0059171597633136, "percentage": 50.3, "elapsed_time": "0:22:09", "remaining_time": "0:21:54"}
+{"current_steps": 180, "total_steps": 338, "loss": 0.4984, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.475302346296336e-05, "epoch": 1.0650887573964498, "percentage": 53.25, "elapsed_time": "0:23:20", "remaining_time": "0:20:29"}
+{"current_steps": 190, "total_steps": 338, "loss": 0.4224, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.2288558050064367e-05, "epoch": 1.1242603550295858, "percentage": 56.21, "elapsed_time": "0:24:30", "remaining_time": "0:19:05"}
+{"current_steps": 200, "total_steps": 338, "loss": 0.4722, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9850534531472546e-05, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:25:40", "remaining_time": "0:17:42"}
+{"current_steps": 200, "total_steps": 338, "loss": null, "eval_loss": 0.496601402759552, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:25:40", "remaining_time": "0:17:42"}
+{"current_steps": 210, "total_steps": 338, "loss": 0.3837, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.746272843378493e-05, "epoch": 1.242603550295858, "percentage": 62.13, "elapsed_time": "0:28:54", "remaining_time": "0:17:37"}
+{"current_steps": 220, "total_steps": 338, "loss": 0.4291, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.5148425564932084e-05, "epoch": 1.301775147928994, "percentage": 65.09, "elapsed_time": "0:30:04", "remaining_time": "0:16:07"}
+{"current_steps": 230, "total_steps": 338, "loss": 0.4442, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.2930194931731382e-05, "epoch": 1.3609467455621302, "percentage": 68.05, "elapsed_time": "0:31:14", "remaining_time": "0:14:40"}
+{"current_steps": 240, "total_steps": 338, "loss": 0.4078, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.0829668647661559e-05, "epoch": 1.4201183431952662, "percentage": 71.01, "elapsed_time": "0:32:23", "remaining_time": "0:13:13"}
+{"current_steps": 250, "total_steps": 338, "loss": 0.4382, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.867330977190877e-06, "epoch": 1.4792899408284024, "percentage": 73.96, "elapsed_time": "0:33:32", "remaining_time": "0:11:48"}
+{"current_steps": 260, "total_steps": 338, "loss": 0.4413, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.062318573891716e-06, "epoch": 1.5384615384615383, "percentage": 76.92, "elapsed_time": "0:34:42", "remaining_time": "0:10:24"}
+{"current_steps": 270, "total_steps": 338, "loss": 0.4165, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.4322338604131715e-06, "epoch": 1.5976331360946747, "percentage": 79.88, "elapsed_time": "0:35:52", "remaining_time": "0:09:02"}
+{"current_steps": 280, "total_steps": 338, "loss": 0.3893, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.992973370223896e-06, "epoch": 1.6568047337278107, "percentage": 82.84, "elapsed_time": "0:37:03", "remaining_time": "0:07:40"}
+{"current_steps": 290, "total_steps": 338, "loss": 0.3823, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.75857272513132e-06, "epoch": 1.7159763313609466, "percentage": 85.8, "elapsed_time": "0:38:13", "remaining_time": "0:06:19"}
+{"current_steps": 300, "total_steps": 338, "loss": 0.4785, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.7410697603511383e-06, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:39:22", "remaining_time": "0:04:59"}
+{"current_steps": 300, "total_steps": 338, "loss": null, "eval_loss": 0.5003606081008911, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:39:22", "remaining_time": "0:04:59"}
+{"current_steps": 310, "total_steps": 338, "loss": 0.3872, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.503871319271551e-07, "epoch": 1.834319526627219, "percentage": 91.72, "elapsed_time": "0:42:36", "remaining_time": "0:03:50"}
+{"current_steps": 320, "total_steps": 338, "loss": 0.4066, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.9423555131007925e-07, "epoch": 1.893491124260355, "percentage": 94.67, "elapsed_time": "0:43:46", "remaining_time": "0:02:27"}
+{"current_steps": 330, "total_steps": 338, "loss": 0.445, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.803859074854425e-08, "epoch": 1.952662721893491, "percentage": 97.63, "elapsed_time": "0:44:56", "remaining_time": "0:01:05"}
+{"current_steps": 338, "total_steps": 338, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:45:53", "remaining_time": "0:00:00"}
+{"current_steps": 38, "total_steps": 38, "loss": null, "eval_loss": 0.496601402759552, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:47:55", "remaining_time": "0:00:00"}

qwen72B_models/qwen_72B_d1_iter2_model/trainer_state.json

@@ -0,0 +1,285 @@
+{
+  "best_metric": 0.496601402759552,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/on_base_iter2_2epoch_d1/checkpoint-200",
+  "epoch": 2.0,
+  "eval_steps": 100,
+  "global_step": 338,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05917159763313609,
+      "grad_norm": 2.225852178863569,
+      "learning_rate": 2.5e-05,
+      "loss": 1.4138,
+      "step": 10
+    },
+    {
+      "epoch": 0.11834319526627218,
+      "grad_norm": 1.6808863792477002,
+      "learning_rate": 5e-05,
+      "loss": 1.0279,
+      "step": 20
+    },
+    {
+      "epoch": 0.17751479289940827,
+      "grad_norm": 0.8652169116376939,
+      "learning_rate": 4.98781004037916e-05,
+      "loss": 0.6745,
+      "step": 30
+    },
+    {
+      "epoch": 0.23668639053254437,
+      "grad_norm": 0.628188364493743,
+      "learning_rate": 4.951359037609088e-05,
+      "loss": 0.5665,
+      "step": 40
+    },
+    {
+      "epoch": 0.2958579881656805,
+      "grad_norm": 2.0643988476119106,
+      "learning_rate": 4.891002460691306e-05,
+      "loss": 0.6138,
+      "step": 50
+    },
+    {
+      "epoch": 0.35502958579881655,
+      "grad_norm": 0.6124814192954576,
+      "learning_rate": 4.807328905014201e-05,
+      "loss": 0.5885,
+      "step": 60
+    },
+    {
+      "epoch": 0.41420118343195267,
+      "grad_norm": 0.9101983521871541,
+      "learning_rate": 4.7011543523897996e-05,
+      "loss": 0.5096,
+      "step": 70
+    },
+    {
+      "epoch": 0.47337278106508873,
+      "grad_norm": 1.2012533572581314,
+      "learning_rate": 4.573514213625505e-05,
+      "loss": 0.5879,
+      "step": 80
+    },
+    {
+      "epoch": 0.5325443786982249,
+      "grad_norm": 0.6264642597880182,
+      "learning_rate": 4.425653231231344e-05,
+      "loss": 0.4712,
+      "step": 90
+    },
+    {
+      "epoch": 0.591715976331361,
+      "grad_norm": 0.8471338152875086,
+      "learning_rate": 4.259013340731224e-05,
+      "loss": 0.4598,
+      "step": 100
+    },
+    {
+      "epoch": 0.591715976331361,
+      "eval_loss": 0.5029934048652649,
+      "eval_runtime": 44.5942,
+      "eval_samples_per_second": 6.727,
+      "eval_steps_per_second": 0.852,
+      "step": 100
+    },
+    {
+      "epoch": 0.650887573964497,
+      "grad_norm": 0.411559425150941,
+      "learning_rate": 4.075219608954278e-05,
+      "loss": 0.5501,
+      "step": 110
+    },
+    {
+      "epoch": 0.7100591715976331,
+      "grad_norm": 0.9138524784632216,
+      "learning_rate": 3.876064386435646e-05,
+      "loss": 0.4926,
+      "step": 120
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": 1.0791585707661275,
+      "learning_rate": 3.663489828471953e-05,
+      "loss": 0.5475,
+      "step": 130
+    },
+    {
+      "epoch": 0.8284023668639053,
+      "grad_norm": 0.7305917551767582,
+      "learning_rate": 3.4395689552855955e-05,
+      "loss": 0.4966,
+      "step": 140
+    },
+    {
+      "epoch": 0.8875739644970414,
+      "grad_norm": 0.9919268654883429,
+      "learning_rate": 3.206485435998498e-05,
+      "loss": 0.5472,
+      "step": 150
+    },
+    {
+      "epoch": 0.9467455621301775,
+      "grad_norm": 1.1793013806565724,
+      "learning_rate": 2.9665122935613727e-05,
+      "loss": 0.5468,
+      "step": 160
+    },
+    {
+      "epoch": 1.0059171597633136,
+      "grad_norm": 0.46458804985635493,
+      "learning_rate": 2.7219897383073373e-05,
+      "loss": 0.5135,
+      "step": 170
+    },
+    {
+      "epoch": 1.0650887573964498,
+      "grad_norm": 0.8719803044635793,
+      "learning_rate": 2.475302346296336e-05,
+      "loss": 0.4984,
+      "step": 180
+    },
+    {
+      "epoch": 1.1242603550295858,
+      "grad_norm": 0.6064445823161934,
+      "learning_rate": 2.2288558050064367e-05,
+      "loss": 0.4224,
+      "step": 190
+    },
+    {
+      "epoch": 1.183431952662722,
+      "grad_norm": 0.5471593713038789,
+      "learning_rate": 1.9850534531472546e-05,
+      "loss": 0.4722,
+      "step": 200
+    },
+    {
+      "epoch": 1.183431952662722,
+      "eval_loss": 0.496601402759552,
+      "eval_runtime": 43.0931,
+      "eval_samples_per_second": 6.962,
+      "eval_steps_per_second": 0.882,
+      "step": 200
+    },
+    {
+      "epoch": 1.242603550295858,
+      "grad_norm": 0.30526160113149914,
+      "learning_rate": 1.746272843378493e-05,
+      "loss": 0.3837,
+      "step": 210
+    },
+    {
+      "epoch": 1.301775147928994,
+      "grad_norm": 1.6844808599149244,
+      "learning_rate": 1.5148425564932084e-05,
+      "loss": 0.4291,
+      "step": 220
+    },
+    {
+      "epoch": 1.3609467455621302,
+      "grad_norm": 0.5210359677169932,
+      "learning_rate": 1.2930194931731382e-05,
+      "loss": 0.4442,
+      "step": 230
+    },
+    {
+      "epoch": 1.4201183431952662,
+      "grad_norm": 1.3407102333836534,
+      "learning_rate": 1.0829668647661559e-05,
+      "loss": 0.4078,
+      "step": 240
+    },
+    {
+      "epoch": 1.4792899408284024,
+      "grad_norm": 0.5141821623131554,
+      "learning_rate": 8.867330977190877e-06,
+      "loss": 0.4382,
+      "step": 250
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": 0.8810333664830745,
+      "learning_rate": 7.062318573891716e-06,
+      "loss": 0.4413,
+      "step": 260
+    },
+    {
+      "epoch": 1.5976331360946747,
+      "grad_norm": 0.9623413258765194,
+      "learning_rate": 5.4322338604131715e-06,
+      "loss": 0.4165,
+      "step": 270
+    },
+    {
+      "epoch": 1.6568047337278107,
+      "grad_norm": 0.707576524188434,
+      "learning_rate": 3.992973370223896e-06,
+      "loss": 0.3893,
+      "step": 280
+    },
+    {
+      "epoch": 1.7159763313609466,
+      "grad_norm": 0.6463531212587023,
+      "learning_rate": 2.75857272513132e-06,
+      "loss": 0.3823,
+      "step": 290
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "grad_norm": 2.502790139467474,
+      "learning_rate": 1.7410697603511383e-06,
+      "loss": 0.4785,
+      "step": 300
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "eval_loss": 0.5003606081008911,
+      "eval_runtime": 42.8999,
+      "eval_samples_per_second": 6.993,
+      "eval_steps_per_second": 0.886,
+      "step": 300
+    },
+    {
+      "epoch": 1.834319526627219,
+      "grad_norm": 0.9588049417813155,
+      "learning_rate": 9.503871319271551e-07,
+      "loss": 0.3872,
+      "step": 310
+    },
+    {
+      "epoch": 1.893491124260355,
+      "grad_norm": 0.7018384917963102,
+      "learning_rate": 3.9423555131007925e-07,
+      "loss": 0.4066,
+      "step": 320
+    },
+    {
+      "epoch": 1.952662721893491,
+      "grad_norm": 1.069849640134778,
+      "learning_rate": 7.803859074854425e-08,
+      "loss": 0.445,
+      "step": 330
+    },
+    {
+      "epoch": 2.0,
+      "step": 338,
+      "total_flos": 312432690462720.0,
+      "train_loss": 0.5284144356405946,
+      "train_runtime": 2754.0946,
+      "train_samples_per_second": 1.961,
+      "train_steps_per_second": 0.123
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 338,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "total_flos": 312432690462720.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen72B_models/qwen_72B_d1_iter2_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13c2fa1d4e480ec4d463516e3353cb0edb779e0f235ef8b4d2a8e1e7cfa3fbbd
+size 6904

qwen72B_models/qwen_72B_d2_iter3_model/README.md

@@ -0,0 +1,69 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B
+model-index:
+- name: on_base_iter3_2epoch_d2_on_base
+  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. -->
+
+# on_base_iter3_2epoch_d2_on_base
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B) on the on_base_iter3_2epoch_d2_on_base dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.4287
+
+## 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: 5e-05
+- train_batch_size: 1
+- eval_batch_size: 1
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 16
+- total_eval_batch_size: 8
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 20
+- num_epochs: 2.0
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.4121        | 0.5917 | 100  | 0.4354          |
+| 0.3406        | 1.1834 | 200  | 0.4287          |
+| 0.3047        | 1.7751 | 300  | 0.4382          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.40.1
+- Pytorch 2.3.0+cu121
+- Datasets 2.19.0
+- Tokenizers 0.19.1

qwen72B_models/qwen_72B_d2_iter3_model/adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/Qwen_72B",
+  "bias": "none",
+  "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": 16,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "o_proj",
+    "q_proj",
+    "down_proj",
+    "up_proj",
+    "k_proj",
+    "gate_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen72B_models/qwen_72B_d2_iter3_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d64f27df1b50fc2d998f6bfeedb15478247bdf13c44f6ad3cfad0b1990e70c9a
+size 209865840

qwen72B_models/qwen_72B_d2_iter3_model/added_tokens.json

@@ -0,0 +1,5 @@
+{
+  "<|endoftext|>": 151643,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644
+}

qwen72B_models/qwen_72B_d2_iter3_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.4286712408065796,
+    "eval_runtime": 42.5309,
+    "eval_samples_per_second": 7.054,
+    "eval_steps_per_second": 0.893,
+    "total_flos": 293955502604288.0,
+    "train_loss": 0.4355149036328468,
+    "train_runtime": 2715.6817,
+    "train_samples_per_second": 1.988,
+    "train_steps_per_second": 0.124
+}

qwen72B_models/qwen_72B_d2_iter3_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.4286712408065796,
+    "eval_runtime": 42.5309,
+    "eval_samples_per_second": 7.054,
+    "eval_steps_per_second": 0.893
+}

qwen72B_models/qwen_72B_d2_iter3_model/special_tokens_map.json

@@ -0,0 +1,27 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

qwen72B_models/qwen_72B_d2_iter3_model/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|endoftext|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{% set system_message = 'You are a helpful assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<|im_start|>system\\n' + system_message + '<|im_end|>\\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|im_start|>user\\n' + content + '<|im_end|>\\n<|im_start|>assistant\\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|endoftext|>' + '\\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "model_max_length": 32768,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

qwen72B_models/qwen_72B_d2_iter3_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 293955502604288.0,
+    "train_loss": 0.4355149036328468,
+    "train_runtime": 2715.6817,
+    "train_samples_per_second": 1.988,
+    "train_steps_per_second": 0.124
+}

qwen72B_models/qwen_72B_d2_iter3_model/trainer_log.jsonl

@@ -0,0 +1,38 @@
+{"current_steps": 10, "total_steps": 338, "loss": 1.5354, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.05917159763313609, "percentage": 2.96, "elapsed_time": "0:01:19", "remaining_time": "0:43:41"}
+{"current_steps": 20, "total_steps": 338, "loss": 1.1732, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.11834319526627218, "percentage": 5.92, "elapsed_time": "0:02:28", "remaining_time": "0:39:22"}
+{"current_steps": 30, "total_steps": 338, "loss": 0.5254, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.98781004037916e-05, "epoch": 0.17751479289940827, "percentage": 8.88, "elapsed_time": "0:03:37", "remaining_time": "0:37:16"}
+{"current_steps": 40, "total_steps": 338, "loss": 0.4683, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.951359037609088e-05, "epoch": 0.23668639053254437, "percentage": 11.83, "elapsed_time": "0:04:46", "remaining_time": "0:35:36"}
+{"current_steps": 50, "total_steps": 338, "loss": 0.446, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.891002460691306e-05, "epoch": 0.2958579881656805, "percentage": 14.79, "elapsed_time": "0:05:56", "remaining_time": "0:34:11"}
+{"current_steps": 60, "total_steps": 338, "loss": 0.5325, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.807328905014201e-05, "epoch": 0.35502958579881655, "percentage": 17.75, "elapsed_time": "0:07:05", "remaining_time": "0:32:52"}
+{"current_steps": 70, "total_steps": 338, "loss": 0.4152, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.7011543523897996e-05, "epoch": 0.41420118343195267, "percentage": 20.71, "elapsed_time": "0:08:14", "remaining_time": "0:31:33"}
+{"current_steps": 80, "total_steps": 338, "loss": 0.408, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.573514213625505e-05, "epoch": 0.47337278106508873, "percentage": 23.67, "elapsed_time": "0:09:24", "remaining_time": "0:30:20"}
+{"current_steps": 90, "total_steps": 338, "loss": 0.4522, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.425653231231344e-05, "epoch": 0.5325443786982249, "percentage": 26.63, "elapsed_time": "0:10:33", "remaining_time": "0:29:05"}
+{"current_steps": 100, "total_steps": 338, "loss": 0.4121, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.259013340731224e-05, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:11:42", "remaining_time": "0:27:50"}
+{"current_steps": 100, "total_steps": 338, "loss": null, "eval_loss": 0.4353737533092499, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:11:42", "remaining_time": "0:27:50"}
+{"current_steps": 110, "total_steps": 338, "loss": 0.4554, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.075219608954278e-05, "epoch": 0.650887573964497, "percentage": 32.54, "elapsed_time": "0:14:55", "remaining_time": "0:30:55"}
+{"current_steps": 120, "total_steps": 338, "loss": 0.386, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.876064386435646e-05, "epoch": 0.7100591715976331, "percentage": 35.5, "elapsed_time": "0:16:04", "remaining_time": "0:29:12"}
+{"current_steps": 130, "total_steps": 338, "loss": 0.3668, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.663489828471953e-05, "epoch": 0.7692307692307693, "percentage": 38.46, "elapsed_time": "0:17:13", "remaining_time": "0:27:34"}
+{"current_steps": 140, "total_steps": 338, "loss": 0.4085, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.4395689552855955e-05, "epoch": 0.8284023668639053, "percentage": 41.42, "elapsed_time": "0:18:23", "remaining_time": "0:25:59"}
+{"current_steps": 150, "total_steps": 338, "loss": 0.4499, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.206485435998498e-05, "epoch": 0.8875739644970414, "percentage": 44.38, "elapsed_time": "0:19:32", "remaining_time": "0:24:29"}
+{"current_steps": 160, "total_steps": 338, "loss": 0.456, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.9665122935613727e-05, "epoch": 0.9467455621301775, "percentage": 47.34, "elapsed_time": "0:20:41", "remaining_time": "0:23:01"}
+{"current_steps": 170, "total_steps": 338, "loss": 0.4702, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.7219897383073373e-05, "epoch": 1.0059171597633136, "percentage": 50.3, "elapsed_time": "0:21:51", "remaining_time": "0:21:35"}
+{"current_steps": 180, "total_steps": 338, "loss": 0.3703, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.475302346296336e-05, "epoch": 1.0650887573964498, "percentage": 53.25, "elapsed_time": "0:23:00", "remaining_time": "0:20:11"}
+{"current_steps": 190, "total_steps": 338, "loss": 0.296, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.2288558050064367e-05, "epoch": 1.1242603550295858, "percentage": 56.21, "elapsed_time": "0:24:09", "remaining_time": "0:18:49"}
+{"current_steps": 200, "total_steps": 338, "loss": 0.3406, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9850534531472546e-05, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:25:18", "remaining_time": "0:17:27"}
+{"current_steps": 200, "total_steps": 338, "loss": null, "eval_loss": 0.4286712408065796, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:25:18", "remaining_time": "0:17:27"}
+{"current_steps": 210, "total_steps": 338, "loss": 0.3035, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.746272843378493e-05, "epoch": 1.242603550295858, "percentage": 62.13, "elapsed_time": "0:28:28", "remaining_time": "0:17:21"}
+{"current_steps": 220, "total_steps": 338, "loss": 0.3362, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.5148425564932084e-05, "epoch": 1.301775147928994, "percentage": 65.09, "elapsed_time": "0:29:37", "remaining_time": "0:15:53"}
+{"current_steps": 230, "total_steps": 338, "loss": 0.3269, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.2930194931731382e-05, "epoch": 1.3609467455621302, "percentage": 68.05, "elapsed_time": "0:30:46", "remaining_time": "0:14:27"}
+{"current_steps": 240, "total_steps": 338, "loss": 0.3322, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.0829668647661559e-05, "epoch": 1.4201183431952662, "percentage": 71.01, "elapsed_time": "0:31:55", "remaining_time": "0:13:01"}
+{"current_steps": 250, "total_steps": 338, "loss": 0.342, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.867330977190877e-06, "epoch": 1.4792899408284024, "percentage": 73.96, "elapsed_time": "0:33:03", "remaining_time": "0:11:38"}
+{"current_steps": 260, "total_steps": 338, "loss": 0.3554, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.062318573891716e-06, "epoch": 1.5384615384615383, "percentage": 76.92, "elapsed_time": "0:34:12", "remaining_time": "0:10:15"}
+{"current_steps": 270, "total_steps": 338, "loss": 0.2876, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.4322338604131715e-06, "epoch": 1.5976331360946747, "percentage": 79.88, "elapsed_time": "0:35:21", "remaining_time": "0:08:54"}
+{"current_steps": 280, "total_steps": 338, "loss": 0.2842, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.992973370223896e-06, "epoch": 1.6568047337278107, "percentage": 82.84, "elapsed_time": "0:36:30", "remaining_time": "0:07:33"}
+{"current_steps": 290, "total_steps": 338, "loss": 0.3026, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.75857272513132e-06, "epoch": 1.7159763313609466, "percentage": 85.8, "elapsed_time": "0:37:39", "remaining_time": "0:06:14"}
+{"current_steps": 300, "total_steps": 338, "loss": 0.3047, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.7410697603511383e-06, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:38:48", "remaining_time": "0:04:54"}
+{"current_steps": 300, "total_steps": 338, "loss": null, "eval_loss": 0.43823307752609253, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:38:48", "remaining_time": "0:04:54"}
+{"current_steps": 310, "total_steps": 338, "loss": 0.377, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.503871319271551e-07, "epoch": 1.834319526627219, "percentage": 91.72, "elapsed_time": "0:42:01", "remaining_time": "0:03:47"}
+{"current_steps": 320, "total_steps": 338, "loss": 0.295, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.9423555131007925e-07, "epoch": 1.893491124260355, "percentage": 94.67, "elapsed_time": "0:43:09", "remaining_time": "0:02:25"}
+{"current_steps": 330, "total_steps": 338, "loss": 0.2967, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.803859074854425e-08, "epoch": 1.952662721893491, "percentage": 97.63, "elapsed_time": "0:44:19", "remaining_time": "0:01:04"}
+{"current_steps": 338, "total_steps": 338, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:45:14", "remaining_time": "0:00:00"}
+{"current_steps": 38, "total_steps": 38, "loss": null, "eval_loss": 0.4286712408065796, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:47:16", "remaining_time": "0:00:00"}

qwen72B_models/qwen_72B_d2_iter3_model/trainer_state.json

@@ -0,0 +1,285 @@
+{
+  "best_metric": 0.4286712408065796,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/models/on_base_iter3_2epoch_d2_on_base/checkpoint-200",
+  "epoch": 2.0,
+  "eval_steps": 100,
+  "global_step": 338,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05917159763313609,
+      "grad_norm": 2.300492870289444,
+      "learning_rate": 2.5e-05,
+      "loss": 1.5354,
+      "step": 10
+    },
+    {
+      "epoch": 0.11834319526627218,
+      "grad_norm": 2.0447049309244134,
+      "learning_rate": 5e-05,
+      "loss": 1.1732,
+      "step": 20
+    },
+    {
+      "epoch": 0.17751479289940827,
+      "grad_norm": 1.4128674057593302,
+      "learning_rate": 4.98781004037916e-05,
+      "loss": 0.5254,
+      "step": 30
+    },
+    {
+      "epoch": 0.23668639053254437,
+      "grad_norm": 1.8437719121139557,
+      "learning_rate": 4.951359037609088e-05,
+      "loss": 0.4683,
+      "step": 40
+    },
+    {
+      "epoch": 0.2958579881656805,
+      "grad_norm": 1.0196903901911931,
+      "learning_rate": 4.891002460691306e-05,
+      "loss": 0.446,
+      "step": 50
+    },
+    {
+      "epoch": 0.35502958579881655,
+      "grad_norm": 0.5793539768141025,
+      "learning_rate": 4.807328905014201e-05,
+      "loss": 0.5325,
+      "step": 60
+    },
+    {
+      "epoch": 0.41420118343195267,
+      "grad_norm": 2.8221048171075607,
+      "learning_rate": 4.7011543523897996e-05,
+      "loss": 0.4152,
+      "step": 70
+    },
+    {
+      "epoch": 0.47337278106508873,
+      "grad_norm": 0.5379795081555752,
+      "learning_rate": 4.573514213625505e-05,
+      "loss": 0.408,
+      "step": 80
+    },
+    {
+      "epoch": 0.5325443786982249,
+      "grad_norm": 0.9173085764985736,
+      "learning_rate": 4.425653231231344e-05,
+      "loss": 0.4522,
+      "step": 90
+    },
+    {
+      "epoch": 0.591715976331361,
+      "grad_norm": 0.6226376725619192,
+      "learning_rate": 4.259013340731224e-05,
+      "loss": 0.4121,
+      "step": 100
+    },
+    {
+      "epoch": 0.591715976331361,
+      "eval_loss": 0.4353737533092499,
+      "eval_runtime": 44.2805,
+      "eval_samples_per_second": 6.775,
+      "eval_steps_per_second": 0.858,
+      "step": 100
+    },
+    {
+      "epoch": 0.650887573964497,
+      "grad_norm": 1.577345855667434,
+      "learning_rate": 4.075219608954278e-05,
+      "loss": 0.4554,
+      "step": 110
+    },
+    {
+      "epoch": 0.7100591715976331,
+      "grad_norm": 0.8169498403794336,
+      "learning_rate": 3.876064386435646e-05,
+      "loss": 0.386,
+      "step": 120
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": 0.2568378898808812,
+      "learning_rate": 3.663489828471953e-05,
+      "loss": 0.3668,
+      "step": 130
+    },
+    {
+      "epoch": 0.8284023668639053,
+      "grad_norm": 0.7731323816903508,
+      "learning_rate": 3.4395689552855955e-05,
+      "loss": 0.4085,
+      "step": 140
+    },
+    {
+      "epoch": 0.8875739644970414,
+      "grad_norm": 0.3275156930124159,
+      "learning_rate": 3.206485435998498e-05,
+      "loss": 0.4499,
+      "step": 150
+    },
+    {
+      "epoch": 0.9467455621301775,
+      "grad_norm": 1.232774744195649,
+      "learning_rate": 2.9665122935613727e-05,
+      "loss": 0.456,
+      "step": 160
+    },
+    {
+      "epoch": 1.0059171597633136,
+      "grad_norm": 0.5225615007237162,
+      "learning_rate": 2.7219897383073373e-05,
+      "loss": 0.4702,
+      "step": 170
+    },
+    {
+      "epoch": 1.0650887573964498,
+      "grad_norm": 0.4682979271736304,
+      "learning_rate": 2.475302346296336e-05,
+      "loss": 0.3703,
+      "step": 180
+    },
+    {
+      "epoch": 1.1242603550295858,
+      "grad_norm": 0.4422825051971799,
+      "learning_rate": 2.2288558050064367e-05,
+      "loss": 0.296,
+      "step": 190
+    },
+    {
+      "epoch": 1.183431952662722,
+      "grad_norm": 0.4943526914540028,
+      "learning_rate": 1.9850534531472546e-05,
+      "loss": 0.3406,
+      "step": 200
+    },
+    {
+      "epoch": 1.183431952662722,
+      "eval_loss": 0.4286712408065796,
+      "eval_runtime": 42.3083,
+      "eval_samples_per_second": 7.091,
+      "eval_steps_per_second": 0.898,
+      "step": 200
+    },
+    {
+      "epoch": 1.242603550295858,
+      "grad_norm": 1.0240142789380449,
+      "learning_rate": 1.746272843378493e-05,
+      "loss": 0.3035,
+      "step": 210
+    },
+    {
+      "epoch": 1.301775147928994,
+      "grad_norm": 0.49822453889115864,
+      "learning_rate": 1.5148425564932084e-05,
+      "loss": 0.3362,
+      "step": 220
+    },
+    {
+      "epoch": 1.3609467455621302,
+      "grad_norm": 1.0897457522370881,
+      "learning_rate": 1.2930194931731382e-05,
+      "loss": 0.3269,
+      "step": 230
+    },
+    {
+      "epoch": 1.4201183431952662,
+      "grad_norm": 0.6236768853827476,
+      "learning_rate": 1.0829668647661559e-05,
+      "loss": 0.3322,
+      "step": 240
+    },
+    {
+      "epoch": 1.4792899408284024,
+      "grad_norm": 0.923548043152688,
+      "learning_rate": 8.867330977190877e-06,
+      "loss": 0.342,
+      "step": 250
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": 0.6810429786190324,
+      "learning_rate": 7.062318573891716e-06,
+      "loss": 0.3554,
+      "step": 260
+    },
+    {
+      "epoch": 1.5976331360946747,
+      "grad_norm": 0.8279379211396576,
+      "learning_rate": 5.4322338604131715e-06,
+      "loss": 0.2876,
+      "step": 270
+    },
+    {
+      "epoch": 1.6568047337278107,
+      "grad_norm": 0.7159229134320508,
+      "learning_rate": 3.992973370223896e-06,
+      "loss": 0.2842,
+      "step": 280
+    },
+    {
+      "epoch": 1.7159763313609466,
+      "grad_norm": 0.5768897528504877,
+      "learning_rate": 2.75857272513132e-06,
+      "loss": 0.3026,
+      "step": 290
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "grad_norm": 0.8681785215993786,
+      "learning_rate": 1.7410697603511383e-06,
+      "loss": 0.3047,
+      "step": 300
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "eval_loss": 0.43823307752609253,
+      "eval_runtime": 42.7053,
+      "eval_samples_per_second": 7.025,
+      "eval_steps_per_second": 0.89,
+      "step": 300
+    },
+    {
+      "epoch": 1.834319526627219,
+      "grad_norm": 0.8956886403019395,
+      "learning_rate": 9.503871319271551e-07,
+      "loss": 0.377,
+      "step": 310
+    },
+    {
+      "epoch": 1.893491124260355,
+      "grad_norm": 0.8581862289502976,
+      "learning_rate": 3.9423555131007925e-07,
+      "loss": 0.295,
+      "step": 320
+    },
+    {
+      "epoch": 1.952662721893491,
+      "grad_norm": 0.903621197764352,
+      "learning_rate": 7.803859074854425e-08,
+      "loss": 0.2967,
+      "step": 330
+    },
+    {
+      "epoch": 2.0,
+      "step": 338,
+      "total_flos": 293955502604288.0,
+      "train_loss": 0.4355149036328468,
+      "train_runtime": 2715.6817,
+      "train_samples_per_second": 1.988,
+      "train_steps_per_second": 0.124
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 338,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "total_flos": 293955502604288.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen72B_models/qwen_72B_d2_iter3_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e63fa59afc8c3d9b50fb2664ed94a668eb98cce9f1ba5bdbc2775688e71e976
+size 6904