Upload 117 files

qwen4B_models/qwen_4B_d0_iter1_model/README.md

@@ -0,0 +1,67 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B
+model-index:
+- name: qwen_4B_d0_iter1_model
+  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. -->
+
+# qwen_4B_d0_iter1_model
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B) on the qwen_4B_d0_iter1_model dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.9073
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.8416        | 1.9608 | 100  | 0.9073          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.41.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

qwen4B_models/qwen_4B_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/models/Qwen1_5_4B",
+  "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": [
+    "v_proj",
+    "down_proj",
+    "k_proj",
+    "gate_proj",
+    "up_proj",
+    "q_proj",
+    "o_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen4B_models/qwen_4B_d0_iter1_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8175c84e0e918587940396bf30b208103503e4658dd185b843cd045f35004eea
+size 31367672

qwen4B_models/qwen_4B_d0_iter1_model/added_tokens.json

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

qwen4B_models/qwen_4B_d0_iter1_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.9073211550712585,
+    "eval_runtime": 5.944,
+    "eval_samples_per_second": 15.141,
+    "eval_steps_per_second": 2.019,
+    "total_flos": 22867310280704.0,
+    "train_loss": 1.0144876452053295,
+    "train_runtime": 349.3742,
+    "train_samples_per_second": 4.631,
+    "train_steps_per_second": 0.292
+}

qwen4B_models/qwen_4B_d0_iter1_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.9073211550712585,
+    "eval_runtime": 5.944,
+    "eval_samples_per_second": 15.141,
+    "eval_steps_per_second": 2.019
+}

qwen4B_models/qwen_4B_d0_iter1_model/merge.log

@@ -0,0 +1,7 @@
+[2024-07-15 08:06:57,661] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+07/15/2024 08:06:59 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
+07/15/2024 08:06:59 - INFO - llmtuner.model.patcher - Using KV cache for faster generation.
+07/15/2024 08:07:00 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
+07/15/2024 08:07:03 - INFO - llmtuner.model.adapter - Merged 1 adapter(s).
+07/15/2024 08:07:03 - INFO - llmtuner.model.adapter - Loaded adapter(s): /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d0_iter1_model
+07/15/2024 08:07:03 - INFO - llmtuner.model.loader - all params: 3950369280

qwen4B_models/qwen_4B_d0_iter1_model/runs/Jul15_07-59-39_t-20240715144706-mwmm8-worker-0/events.out.tfevents.1721030446.t-20240715144706-mwmm8-worker-0.40318.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:944398a9a11683d59b27cd3a49d1dc3f15d334fe687b1d5c723df96bc176a13c
+size 8006

qwen4B_models/qwen_4B_d0_iter1_model/runs/Jul15_07-59-39_t-20240715144706-mwmm8-worker-0/events.out.tfevents.1721030806.t-20240715144706-mwmm8-worker-0.40318.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f56a1f13348f9fafaaabdcae298ba36f54aa16bdec9112769d4b0af92b6b3f3f
+size 354

qwen4B_models/qwen_4B_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
+  }
+}

qwen4B_models/qwen_4B_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
+}

qwen4B_models/qwen_4B_d0_iter1_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 22867310280704.0,
+    "train_loss": 1.0144876452053295,
+    "train_runtime": 349.3742,
+    "train_samples_per_second": 4.631,
+    "train_steps_per_second": 0.292
+}

qwen4B_models/qwen_4B_d0_iter1_model/trainer_log.jsonl

@@ -0,0 +1,13 @@
+{"current_steps": 10, "total_steps": 102, "loss": 1.302, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.19607843137254902, "percentage": 9.8, "elapsed_time": "0:00:48", "remaining_time": "0:07:25"}
+{"current_steps": 20, "total_steps": 102, "loss": 1.2195, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.39215686274509803, "percentage": 19.61, "elapsed_time": "0:01:19", "remaining_time": "0:05:25"}
+{"current_steps": 30, "total_steps": 102, "loss": 0.9886, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.8187561277552374e-05, "epoch": 0.5882352941176471, "percentage": 29.41, "elapsed_time": "0:01:50", "remaining_time": "0:04:24"}
+{"current_steps": 40, "total_steps": 102, "loss": 1.0972, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.301303984001967e-05, "epoch": 0.7843137254901961, "percentage": 39.22, "elapsed_time": "0:02:21", "remaining_time": "0:03:39"}
+{"current_steps": 50, "total_steps": 102, "loss": 0.9888, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.5226715929283506e-05, "epoch": 0.9803921568627451, "percentage": 49.02, "elapsed_time": "0:02:52", "remaining_time": "0:02:59"}
+{"current_steps": 60, "total_steps": 102, "loss": 0.9882, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.595756834225089e-05, "epoch": 1.1764705882352942, "percentage": 58.82, "elapsed_time": "0:03:23", "remaining_time": "0:02:22"}
+{"current_steps": 70, "total_steps": 102, "loss": 0.9159, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.6549578039787436e-05, "epoch": 1.3725490196078431, "percentage": 68.63, "elapsed_time": "0:03:54", "remaining_time": "0:01:47"}
+{"current_steps": 80, "total_steps": 102, "loss": 0.8981, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.36685749586087e-06, "epoch": 1.5686274509803921, "percentage": 78.43, "elapsed_time": "0:04:25", "remaining_time": "0:01:13"}
+{"current_steps": 90, "total_steps": 102, "loss": 0.9455, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.595861075973613e-06, "epoch": 1.7647058823529411, "percentage": 88.24, "elapsed_time": "0:04:56", "remaining_time": "0:00:39"}
+{"current_steps": 100, "total_steps": 102, "loss": 0.8416, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.335497040648898e-08, "epoch": 1.9607843137254903, "percentage": 98.04, "elapsed_time": "0:05:27", "remaining_time": "0:00:06"}
+{"current_steps": 100, "total_steps": 102, "loss": null, "eval_loss": 0.9073211550712585, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.9607843137254903, "percentage": 98.04, "elapsed_time": "0:05:27", "remaining_time": "0:00:06"}
+{"current_steps": 102, "total_steps": 102, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:05:49", "remaining_time": "0:00:00"}
+{"current_steps": 12, "total_steps": 12, "loss": null, "eval_loss": 0.9073211550712585, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:06:00", "remaining_time": "0:00:00"}

qwen4B_models/qwen_4B_d0_iter1_model/trainer_state.json

@@ -0,0 +1,120 @@
+{
+  "best_metric": 0.9073211550712585,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d0_iter1_model/checkpoint-100",
+  "epoch": 2.0,
+  "eval_steps": 100,
+  "global_step": 102,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.19607843137254902,
+      "grad_norm": 1.8514470549546265,
+      "learning_rate": 2.5e-05,
+      "loss": 1.302,
+      "step": 10
+    },
+    {
+      "epoch": 0.39215686274509803,
+      "grad_norm": 0.3635169007023558,
+      "learning_rate": 5e-05,
+      "loss": 1.2195,
+      "step": 20
+    },
+    {
+      "epoch": 0.5882352941176471,
+      "grad_norm": 0.6423674064129872,
+      "learning_rate": 4.8187561277552374e-05,
+      "loss": 0.9886,
+      "step": 30
+    },
+    {
+      "epoch": 0.7843137254901961,
+      "grad_norm": 1.7652163975371618,
+      "learning_rate": 4.301303984001967e-05,
+      "loss": 1.0972,
+      "step": 40
+    },
+    {
+      "epoch": 0.9803921568627451,
+      "grad_norm": 0.8125720365482934,
+      "learning_rate": 3.5226715929283506e-05,
+      "loss": 0.9888,
+      "step": 50
+    },
+    {
+      "epoch": 1.1764705882352942,
+      "grad_norm": 0.5160887085744938,
+      "learning_rate": 2.595756834225089e-05,
+      "loss": 0.9882,
+      "step": 60
+    },
+    {
+      "epoch": 1.3725490196078431,
+      "grad_norm": 0.9475974337041686,
+      "learning_rate": 1.6549578039787436e-05,
+      "loss": 0.9159,
+      "step": 70
+    },
+    {
+      "epoch": 1.5686274509803921,
+      "grad_norm": 0.6992885675683204,
+      "learning_rate": 8.36685749586087e-06,
+      "loss": 0.8981,
+      "step": 80
+    },
+    {
+      "epoch": 1.7647058823529411,
+      "grad_norm": 0.6476094908491726,
+      "learning_rate": 2.595861075973613e-06,
+      "loss": 0.9455,
+      "step": 90
+    },
+    {
+      "epoch": 1.9607843137254903,
+      "grad_norm": 0.852996571624781,
+      "learning_rate": 7.335497040648898e-08,
+      "loss": 0.8416,
+      "step": 100
+    },
+    {
+      "epoch": 1.9607843137254903,
+      "eval_loss": 0.9073211550712585,
+      "eval_runtime": 10.152,
+      "eval_samples_per_second": 8.865,
+      "eval_steps_per_second": 1.182,
+      "step": 100
+    },
+    {
+      "epoch": 2.0,
+      "step": 102,
+      "total_flos": 22867310280704.0,
+      "train_loss": 1.0144876452053295,
+      "train_runtime": 349.3742,
+      "train_samples_per_second": 4.631,
+      "train_steps_per_second": 0.292
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 102,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 22867310280704.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen4B_models/qwen_4B_d0_iter1_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48c5fdf5bb26762d6b5ea66c5dcd14133315ee9223ab03c8e9dc85c8c2c3a758
+size 7032

qwen4B_models/qwen_4B_d1_iter2_model/README.md

@@ -0,0 +1,67 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B
+model-index:
+- name: qwen_4B_d1_iter2_model
+  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. -->
+
+# qwen_4B_d1_iter2_model
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B) on the qwen_4B_d1_iter2_model dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.7857
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.6978        | 1.4706 | 100  | 0.7857          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.41.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

qwen4B_models/qwen_4B_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/models/Qwen1_5_4B",
+  "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": [
+    "k_proj",
+    "down_proj",
+    "o_proj",
+    "up_proj",
+    "q_proj",
+    "gate_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen4B_models/qwen_4B_d1_iter2_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:223f5dc1d3695c84b692cdd8615a694fbe0fef2c71f008057debcec98e041b99
+size 31367672

qwen4B_models/qwen_4B_d1_iter2_model/added_tokens.json

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

qwen4B_models/qwen_4B_d1_iter2_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.7856930494308472,
+    "eval_runtime": 7.9389,
+    "eval_samples_per_second": 15.241,
+    "eval_steps_per_second": 2.015,
+    "total_flos": 21111522983936.0,
+    "train_loss": 0.8778598624117234,
+    "train_runtime": 458.5184,
+    "train_samples_per_second": 4.746,
+    "train_steps_per_second": 0.297
+}

qwen4B_models/qwen_4B_d1_iter2_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.7856930494308472,
+    "eval_runtime": 7.9389,
+    "eval_samples_per_second": 15.241,
+    "eval_steps_per_second": 2.015
+}

qwen4B_models/qwen_4B_d1_iter2_model/merge.log

@@ -0,0 +1,7 @@
+[2024-07-15 10:10:22,449] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+07/15/2024 10:10:24 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
+07/15/2024 10:10:24 - INFO - llmtuner.model.patcher - Using KV cache for faster generation.
+07/15/2024 10:10:25 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
+07/15/2024 10:10:28 - INFO - llmtuner.model.adapter - Merged 1 adapter(s).
+07/15/2024 10:10:28 - INFO - llmtuner.model.adapter - Loaded adapter(s): /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d1_iter2_model
+07/15/2024 10:10:28 - INFO - llmtuner.model.loader - all params: 3950369280

qwen4B_models/qwen_4B_d1_iter2_model/runs/Jul15_10-01-16_t-20240715144706-mwmm8-worker-0/events.out.tfevents.1721037741.t-20240715144706-mwmm8-worker-0.91732.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c04452716c83e238a9ac89ae188b5a3b3b9562d3492997299769ca77c6845b79
+size 8637

qwen4B_models/qwen_4B_d1_iter2_model/runs/Jul15_10-01-16_t-20240715144706-mwmm8-worker-0/events.out.tfevents.1721038213.t-20240715144706-mwmm8-worker-0.91732.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93f38c90c8428f4bf948c6b41ee9f7fc583e9ea3840c292b1df818734fb21620
+size 359

qwen4B_models/qwen_4B_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
+  }
+}

qwen4B_models/qwen_4B_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
+}

qwen4B_models/qwen_4B_d1_iter2_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 21111522983936.0,
+    "train_loss": 0.8778598624117234,
+    "train_runtime": 458.5184,
+    "train_samples_per_second": 4.746,
+    "train_steps_per_second": 0.297
+}

qwen4B_models/qwen_4B_d1_iter2_model/trainer_log.jsonl

@@ -0,0 +1,16 @@
+{"current_steps": 10, "total_steps": 136, "loss": 1.438, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.14705882352941177, "percentage": 7.35, "elapsed_time": "0:00:48", "remaining_time": "0:10:07"}
+{"current_steps": 20, "total_steps": 136, "loss": 1.2455, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.29411764705882354, "percentage": 14.71, "elapsed_time": "0:01:19", "remaining_time": "0:07:41"}
+{"current_steps": 30, "total_steps": 136, "loss": 0.8748, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.908874981298057e-05, "epoch": 0.4411764705882353, "percentage": 22.06, "elapsed_time": "0:01:50", "remaining_time": "0:06:30"}
+{"current_steps": 40, "total_steps": 136, "loss": 0.8399, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.642142940418973e-05, "epoch": 0.5882352941176471, "percentage": 29.41, "elapsed_time": "0:02:21", "remaining_time": "0:05:40"}
+{"current_steps": 50, "total_steps": 136, "loss": 0.8371, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.2192486471335585e-05, "epoch": 0.7352941176470589, "percentage": 36.76, "elapsed_time": "0:02:52", "remaining_time": "0:04:56"}
+{"current_steps": 60, "total_steps": 136, "loss": 0.8346, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.671021101749476e-05, "epoch": 0.8823529411764706, "percentage": 44.12, "elapsed_time": "0:03:23", "remaining_time": "0:04:18"}
+{"current_steps": 70, "total_steps": 136, "loss": 0.85, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.0374261005275607e-05, "epoch": 1.0294117647058822, "percentage": 51.47, "elapsed_time": "0:03:55", "remaining_time": "0:03:41"}
+{"current_steps": 80, "total_steps": 136, "loss": 0.8218, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.3646527285364565e-05, "epoch": 1.1764705882352942, "percentage": 58.82, "elapsed_time": "0:04:26", "remaining_time": "0:03:06"}
+{"current_steps": 90, "total_steps": 136, "loss": 0.8021, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.7017461746600506e-05, "epoch": 1.3235294117647058, "percentage": 66.18, "elapsed_time": "0:04:57", "remaining_time": "0:02:32"}
+{"current_steps": 100, "total_steps": 136, "loss": 0.6978, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.0970323365940444e-05, "epoch": 1.4705882352941178, "percentage": 73.53, "elapsed_time": "0:05:28", "remaining_time": "0:01:58"}
+{"current_steps": 100, "total_steps": 136, "loss": null, "eval_loss": 0.7856930494308472, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.4705882352941178, "percentage": 73.53, "elapsed_time": "0:05:28", "remaining_time": "0:01:58"}
+{"current_steps": 110, "total_steps": 136, "loss": 0.7205, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.945948621809091e-06, "epoch": 1.6176470588235294, "percentage": 80.88, "elapsed_time": "0:06:17", "remaining_time": "0:01:29"}
+{"current_steps": 120, "total_steps": 136, "loss": 0.787, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.310614508226078e-06, "epoch": 1.7647058823529411, "percentage": 88.24, "elapsed_time": "0:06:48", "remaining_time": "0:00:54"}
+{"current_steps": 130, "total_steps": 136, "loss": 0.7046, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.293369364618465e-07, "epoch": 1.9117647058823528, "percentage": 95.59, "elapsed_time": "0:07:19", "remaining_time": "0:00:20"}
+{"current_steps": 136, "total_steps": 136, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:07:38", "remaining_time": "0:00:00"}
+{"current_steps": 16, "total_steps": 16, "loss": null, "eval_loss": 0.7856930494308472, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:07:51", "remaining_time": "0:00:00"}

qwen4B_models/qwen_4B_d1_iter2_model/trainer_state.json

@@ -0,0 +1,141 @@
+{
+  "best_metric": 0.7856930494308472,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d1_iter2_model/checkpoint-100",
+  "epoch": 2.0,
+  "eval_steps": 100,
+  "global_step": 136,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.14705882352941177,
+      "grad_norm": 1.1161658260127314,
+      "learning_rate": 2.5e-05,
+      "loss": 1.438,
+      "step": 10
+    },
+    {
+      "epoch": 0.29411764705882354,
+      "grad_norm": 1.141386114916924,
+      "learning_rate": 5e-05,
+      "loss": 1.2455,
+      "step": 20
+    },
+    {
+      "epoch": 0.4411764705882353,
+      "grad_norm": 0.908747056691058,
+      "learning_rate": 4.908874981298057e-05,
+      "loss": 0.8748,
+      "step": 30
+    },
+    {
+      "epoch": 0.5882352941176471,
+      "grad_norm": 1.160724824332908,
+      "learning_rate": 4.642142940418973e-05,
+      "loss": 0.8399,
+      "step": 40
+    },
+    {
+      "epoch": 0.7352941176470589,
+      "grad_norm": 1.7354204311716015,
+      "learning_rate": 4.2192486471335585e-05,
+      "loss": 0.8371,
+      "step": 50
+    },
+    {
+      "epoch": 0.8823529411764706,
+      "grad_norm": 0.989621687985043,
+      "learning_rate": 3.671021101749476e-05,
+      "loss": 0.8346,
+      "step": 60
+    },
+    {
+      "epoch": 1.0294117647058822,
+      "grad_norm": 1.4467052607165527,
+      "learning_rate": 3.0374261005275607e-05,
+      "loss": 0.85,
+      "step": 70
+    },
+    {
+      "epoch": 1.1764705882352942,
+      "grad_norm": 1.022122556125995,
+      "learning_rate": 2.3646527285364565e-05,
+      "loss": 0.8218,
+      "step": 80
+    },
+    {
+      "epoch": 1.3235294117647058,
+      "grad_norm": 1.168833164765407,
+      "learning_rate": 1.7017461746600506e-05,
+      "loss": 0.8021,
+      "step": 90
+    },
+    {
+      "epoch": 1.4705882352941178,
+      "grad_norm": 0.4995411891529773,
+      "learning_rate": 1.0970323365940444e-05,
+      "loss": 0.6978,
+      "step": 100
+    },
+    {
+      "epoch": 1.4705882352941178,
+      "eval_loss": 0.7856930494308472,
+      "eval_runtime": 12.019,
+      "eval_samples_per_second": 10.067,
+      "eval_steps_per_second": 1.331,
+      "step": 100
+    },
+    {
+      "epoch": 1.6176470588235294,
+      "grad_norm": 1.6403596296455754,
+      "learning_rate": 5.945948621809091e-06,
+      "loss": 0.7205,
+      "step": 110
+    },
+    {
+      "epoch": 1.7647058823529411,
+      "grad_norm": 0.8006724271557235,
+      "learning_rate": 2.310614508226078e-06,
+      "loss": 0.787,
+      "step": 120
+    },
+    {
+      "epoch": 1.9117647058823528,
+      "grad_norm": 1.7764723276974967,
+      "learning_rate": 3.293369364618465e-07,
+      "loss": 0.7046,
+      "step": 130
+    },
+    {
+      "epoch": 2.0,
+      "step": 136,
+      "total_flos": 21111522983936.0,
+      "train_loss": 0.8778598624117234,
+      "train_runtime": 458.5184,
+      "train_samples_per_second": 4.746,
+      "train_steps_per_second": 0.297
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 136,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 21111522983936.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen4B_models/qwen_4B_d1_iter2_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a826000e4020639bf4b260a06a4ff200ad767f6f072c88c333b42fbbd4e4a65
+size 7032

qwen4B_models/qwen_4B_d2_iter3_model/README.md

@@ -0,0 +1,67 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B
+model-index:
+- name: qwen_4B_d2_iter3_model
+  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. -->
+
+# qwen_4B_d2_iter3_model
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_4B) on the qwen_4B_d2_iter3_model dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.7803
+
+## 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: 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 |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.6777        | 1.3072 | 100  | 0.7803          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.41.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

qwen4B_models/qwen_4B_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/models/Qwen1_5_4B",
+  "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",
+    "down_proj",
+    "up_proj",
+    "v_proj",
+    "o_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen4B_models/qwen_4B_d2_iter3_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1e6fa97eb51fb31d6156d36d6468dbadb1a8edc635ecf9219b866b6c42c3291b
+size 31367672

qwen4B_models/qwen_4B_d2_iter3_model/added_tokens.json

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

qwen4B_models/qwen_4B_d2_iter3_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 1.9869281045751634,
+    "eval_loss": 0.7803115844726562,
+    "eval_runtime": 8.3907,
+    "eval_samples_per_second": 16.208,
+    "eval_steps_per_second": 2.026,
+    "total_flos": 20289411874816.0,
+    "train_loss": 0.8229627185746243,
+    "train_runtime": 505.2345,
+    "train_samples_per_second": 4.822,
+    "train_steps_per_second": 0.301
+}

qwen4B_models/qwen_4B_d2_iter3_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 1.9869281045751634,
+    "eval_loss": 0.7803115844726562,
+    "eval_runtime": 8.3907,
+    "eval_samples_per_second": 16.208,
+    "eval_steps_per_second": 2.026
+}

qwen4B_models/qwen_4B_d2_iter3_model/merge.log

@@ -0,0 +1,7 @@
+[2024-07-15 12:18:22,063] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+07/15/2024 12:18:24 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
+07/15/2024 12:18:24 - INFO - llmtuner.model.patcher - Using KV cache for faster generation.
+07/15/2024 12:18:24 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
+07/15/2024 12:18:27 - INFO - llmtuner.model.adapter - Merged 1 adapter(s).
+07/15/2024 12:18:27 - INFO - llmtuner.model.adapter - Loaded adapter(s): /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen4B_models/qwen_4B_d2_iter3_model
+07/15/2024 12:18:27 - INFO - llmtuner.model.loader - all params: 3950369280

qwen4B_models/qwen_4B_d2_iter3_model/runs/Jul15_12-07-19_t-20240715144706-mwmm8-worker-0/events.out.tfevents.1721045373.t-20240715144706-mwmm8-worker-0.144068.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b4992da340f51fafb479efca3bc0c9e5a3a44ce71daf064f0f85e6b335e4286
+size 9059