Upload 168 files

qwen14B_models/qwen_14B_d0_iter1_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/models/Qwen1_5_14B
+model-index:
+- name: qwen_14B_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_14B_d0_iter1_model
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B) on the qwen_14B_d0_iter1_model dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6997
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.6716        | 0.5917 | 100  | 0.7039          |
+| 0.5798        | 1.1834 | 200  | 0.7010          |
+| 0.5292        | 1.7751 | 300  | 0.6997          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.41.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

qwen14B_models/qwen_14B_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_14B",
+  "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": [
+    "gate_proj",
+    "v_proj",
+    "up_proj",
+    "o_proj",
+    "k_proj",
+    "down_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen14B_models/qwen_14B_d0_iter1_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:538f964fab9f9d87ce8a6cad7c3fe279c0a1e4a4af3caa380efebce6c47beba3
+size 62415664

qwen14B_models/qwen_14B_d0_iter1_model/added_tokens.json

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

qwen14B_models/qwen_14B_d0_iter1_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.6997424960136414,
+    "eval_runtime": 18.495,
+    "eval_samples_per_second": 16.221,
+    "eval_steps_per_second": 2.055,
+    "total_flos": 123391664193536.0,
+    "train_loss": 0.6494796699320776,
+    "train_runtime": 1192.6701,
+    "train_samples_per_second": 4.528,
+    "train_steps_per_second": 0.283
+}

qwen14B_models/qwen_14B_d0_iter1_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.6997424960136414,
+    "eval_runtime": 18.495,
+    "eval_samples_per_second": 16.221,
+    "eval_steps_per_second": 2.055
+}

qwen14B_models/qwen_14B_d0_iter1_model/merge.log

@@ -0,0 +1,7 @@
+[2024-07-15 09:01:42,071] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+07/15/2024 09:01:44 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
+07/15/2024 09:01:44 - INFO - llmtuner.model.patcher - Using KV cache for faster generation.
+07/15/2024 09:01:45 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
+07/15/2024 09:01:58 - INFO - llmtuner.model.adapter - Merged 1 adapter(s).
+07/15/2024 09:01:58 - INFO - llmtuner.model.adapter - Loaded adapter(s): /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen14B_models/qwen_14B_d0_iter1_model
+07/15/2024 09:01:58 - INFO - llmtuner.model.loader - all params: 14167290880

qwen14B_models/qwen_14B_d0_iter1_model/runs/Jul15_08-39-17_t-20240715144206-zn57g-worker-0/events.out.tfevents.1721032863.t-20240715144206-zn57g-worker-0.53949.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ae4ec621e45944bac0de053da4552893deafe5696c64a1035b207aed6b2a3fe
+size 13407

qwen14B_models/qwen_14B_d0_iter1_model/runs/Jul15_08-39-17_t-20240715144206-zn57g-worker-0/events.out.tfevents.1721034091.t-20240715144206-zn57g-worker-0.53949.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9361689649b9278c4d6f2cb095d69c32bc3baf38985589cba28eb1800da95a92
+size 359

qwen14B_models/qwen_14B_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
+  }
+}

qwen14B_models/qwen_14B_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
+}

qwen14B_models/qwen_14B_d0_iter1_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 123391664193536.0,
+    "train_loss": 0.6494796699320776,
+    "train_runtime": 1192.6701,
+    "train_samples_per_second": 4.528,
+    "train_steps_per_second": 0.283
+}

qwen14B_models/qwen_14B_d0_iter1_model/trainer_log.jsonl

@@ -0,0 +1,38 @@
+{"current_steps": 10, "total_steps": 338, "loss": 1.0824, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.05917159763313609, "percentage": 2.96, "elapsed_time": "0:00:40", "remaining_time": "0:21:59"}
+{"current_steps": 20, "total_steps": 338, "loss": 1.0109, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.11834319526627218, "percentage": 5.92, "elapsed_time": "0:01:11", "remaining_time": "0:19:01"}
+{"current_steps": 30, "total_steps": 338, "loss": 0.6595, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.98781004037916e-05, "epoch": 0.17751479289940827, "percentage": 8.88, "elapsed_time": "0:01:43", "remaining_time": "0:17:43"}
+{"current_steps": 40, "total_steps": 338, "loss": 0.7391, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.951359037609088e-05, "epoch": 0.23668639053254437, "percentage": 11.83, "elapsed_time": "0:02:15", "remaining_time": "0:16:51"}
+{"current_steps": 50, "total_steps": 338, "loss": 0.7051, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.891002460691306e-05, "epoch": 0.2958579881656805, "percentage": 14.79, "elapsed_time": "0:02:47", "remaining_time": "0:16:03"}
+{"current_steps": 60, "total_steps": 338, "loss": 0.7091, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.807328905014201e-05, "epoch": 0.35502958579881655, "percentage": 17.75, "elapsed_time": "0:03:18", "remaining_time": "0:15:21"}
+{"current_steps": 70, "total_steps": 338, "loss": 0.6642, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.7011543523897996e-05, "epoch": 0.41420118343195267, "percentage": 20.71, "elapsed_time": "0:03:50", "remaining_time": "0:14:44"}
+{"current_steps": 80, "total_steps": 338, "loss": 0.6352, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.573514213625505e-05, "epoch": 0.47337278106508873, "percentage": 23.67, "elapsed_time": "0:04:22", "remaining_time": "0:14:06"}
+{"current_steps": 90, "total_steps": 338, "loss": 0.695, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.425653231231344e-05, "epoch": 0.5325443786982249, "percentage": 26.63, "elapsed_time": "0:04:54", "remaining_time": "0:13:30"}
+{"current_steps": 100, "total_steps": 338, "loss": 0.6716, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.259013340731224e-05, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:05:26", "remaining_time": "0:12:56"}
+{"current_steps": 100, "total_steps": 338, "loss": null, "eval_loss": 0.703880250453949, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:05:26", "remaining_time": "0:12:56"}
+{"current_steps": 110, "total_steps": 338, "loss": 0.6649, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.075219608954278e-05, "epoch": 0.650887573964497, "percentage": 32.54, "elapsed_time": "0:06:34", "remaining_time": "0:13:37"}
+{"current_steps": 120, "total_steps": 338, "loss": 0.6485, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.876064386435646e-05, "epoch": 0.7100591715976331, "percentage": 35.5, "elapsed_time": "0:07:06", "remaining_time": "0:12:54"}
+{"current_steps": 130, "total_steps": 338, "loss": 0.5978, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.663489828471953e-05, "epoch": 0.7692307692307693, "percentage": 38.46, "elapsed_time": "0:07:38", "remaining_time": "0:12:13"}
+{"current_steps": 140, "total_steps": 338, "loss": 0.6504, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.4395689552855955e-05, "epoch": 0.8284023668639053, "percentage": 41.42, "elapsed_time": "0:08:10", "remaining_time": "0:11:33"}
+{"current_steps": 150, "total_steps": 338, "loss": 0.6214, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.206485435998498e-05, "epoch": 0.8875739644970414, "percentage": 44.38, "elapsed_time": "0:08:41", "remaining_time": "0:10:53"}
+{"current_steps": 160, "total_steps": 338, "loss": 0.6144, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.9665122935613727e-05, "epoch": 0.9467455621301775, "percentage": 47.34, "elapsed_time": "0:09:13", "remaining_time": "0:10:15"}
+{"current_steps": 170, "total_steps": 338, "loss": 0.6493, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.7219897383073373e-05, "epoch": 1.0059171597633136, "percentage": 50.3, "elapsed_time": "0:09:45", "remaining_time": "0:09:38"}
+{"current_steps": 180, "total_steps": 338, "loss": 0.6474, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.475302346296336e-05, "epoch": 1.0650887573964498, "percentage": 53.25, "elapsed_time": "0:10:16", "remaining_time": "0:09:01"}
+{"current_steps": 190, "total_steps": 338, "loss": 0.564, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.2288558050064367e-05, "epoch": 1.1242603550295858, "percentage": 56.21, "elapsed_time": "0:10:48", "remaining_time": "0:08:25"}
+{"current_steps": 200, "total_steps": 338, "loss": 0.5798, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9850534531472546e-05, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:11:20", "remaining_time": "0:07:49"}
+{"current_steps": 200, "total_steps": 338, "loss": null, "eval_loss": 0.7010006904602051, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:11:20", "remaining_time": "0:07:49"}
+{"current_steps": 210, "total_steps": 338, "loss": 0.6323, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.746272843378493e-05, "epoch": 1.242603550295858, "percentage": 62.13, "elapsed_time": "0:12:27", "remaining_time": "0:07:35"}
+{"current_steps": 220, "total_steps": 338, "loss": 0.6106, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.5148425564932084e-05, "epoch": 1.301775147928994, "percentage": 65.09, "elapsed_time": "0:12:59", "remaining_time": "0:06:58"}
+{"current_steps": 230, "total_steps": 338, "loss": 0.6374, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.2930194931731382e-05, "epoch": 1.3609467455621302, "percentage": 68.05, "elapsed_time": "0:13:31", "remaining_time": "0:06:21"}
+{"current_steps": 240, "total_steps": 338, "loss": 0.5758, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.0829668647661559e-05, "epoch": 1.4201183431952662, "percentage": 71.01, "elapsed_time": "0:14:03", "remaining_time": "0:05:44"}
+{"current_steps": 250, "total_steps": 338, "loss": 0.5988, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.867330977190877e-06, "epoch": 1.4792899408284024, "percentage": 73.96, "elapsed_time": "0:14:35", "remaining_time": "0:05:08"}
+{"current_steps": 260, "total_steps": 338, "loss": 0.5956, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.062318573891716e-06, "epoch": 1.5384615384615383, "percentage": 76.92, "elapsed_time": "0:15:07", "remaining_time": "0:04:32"}
+{"current_steps": 270, "total_steps": 338, "loss": 0.5467, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.4322338604131715e-06, "epoch": 1.5976331360946747, "percentage": 79.88, "elapsed_time": "0:15:40", "remaining_time": "0:03:56"}
+{"current_steps": 280, "total_steps": 338, "loss": 0.599, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.992973370223896e-06, "epoch": 1.6568047337278107, "percentage": 82.84, "elapsed_time": "0:16:11", "remaining_time": "0:03:21"}
+{"current_steps": 290, "total_steps": 338, "loss": 0.5505, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.75857272513132e-06, "epoch": 1.7159763313609466, "percentage": 85.8, "elapsed_time": "0:16:43", "remaining_time": "0:02:46"}
+{"current_steps": 300, "total_steps": 338, "loss": 0.5292, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.7410697603511383e-06, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:17:15", "remaining_time": "0:02:11"}
+{"current_steps": 300, "total_steps": 338, "loss": null, "eval_loss": 0.6997424960136414, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:17:15", "remaining_time": "0:02:11"}
+{"current_steps": 310, "total_steps": 338, "loss": 0.6355, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.503871319271551e-07, "epoch": 1.834319526627219, "percentage": 91.72, "elapsed_time": "0:18:21", "remaining_time": "0:01:39"}
+{"current_steps": 320, "total_steps": 338, "loss": 0.6309, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.9423555131007925e-07, "epoch": 1.893491124260355, "percentage": 94.67, "elapsed_time": "0:18:53", "remaining_time": "0:01:03"}
+{"current_steps": 330, "total_steps": 338, "loss": 0.5814, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.803859074854425e-08, "epoch": 1.952662721893491, "percentage": 97.63, "elapsed_time": "0:19:25", "remaining_time": "0:00:28"}
+{"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": "0:19:52", "remaining_time": "0:00:00"}
+{"current_steps": 38, "total_steps": 38, "loss": null, "eval_loss": 0.6997424960136414, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:20:27", "remaining_time": "0:00:00"}

qwen14B_models/qwen_14B_d0_iter1_model/trainer_state.json

@@ -0,0 +1,297 @@
+{
+  "best_metric": 0.6997424960136414,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen14B_models/qwen_14B_d0_iter1_model/checkpoint-300",
+  "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": 0.586446509459034,
+      "learning_rate": 2.5e-05,
+      "loss": 1.0824,
+      "step": 10
+    },
+    {
+      "epoch": 0.11834319526627218,
+      "grad_norm": 2.2316997759806494,
+      "learning_rate": 5e-05,
+      "loss": 1.0109,
+      "step": 20
+    },
+    {
+      "epoch": 0.17751479289940827,
+      "grad_norm": 0.4074702236613225,
+      "learning_rate": 4.98781004037916e-05,
+      "loss": 0.6595,
+      "step": 30
+    },
+    {
+      "epoch": 0.23668639053254437,
+      "grad_norm": 0.3126617987747972,
+      "learning_rate": 4.951359037609088e-05,
+      "loss": 0.7391,
+      "step": 40
+    },
+    {
+      "epoch": 0.2958579881656805,
+      "grad_norm": 0.708271236864645,
+      "learning_rate": 4.891002460691306e-05,
+      "loss": 0.7051,
+      "step": 50
+    },
+    {
+      "epoch": 0.35502958579881655,
+      "grad_norm": 0.3826205284015388,
+      "learning_rate": 4.807328905014201e-05,
+      "loss": 0.7091,
+      "step": 60
+    },
+    {
+      "epoch": 0.41420118343195267,
+      "grad_norm": 0.4309917712753102,
+      "learning_rate": 4.7011543523897996e-05,
+      "loss": 0.6642,
+      "step": 70
+    },
+    {
+      "epoch": 0.47337278106508873,
+      "grad_norm": 0.46078253102774147,
+      "learning_rate": 4.573514213625505e-05,
+      "loss": 0.6352,
+      "step": 80
+    },
+    {
+      "epoch": 0.5325443786982249,
+      "grad_norm": 0.350545318021008,
+      "learning_rate": 4.425653231231344e-05,
+      "loss": 0.695,
+      "step": 90
+    },
+    {
+      "epoch": 0.591715976331361,
+      "grad_norm": 1.3918929820024042,
+      "learning_rate": 4.259013340731224e-05,
+      "loss": 0.6716,
+      "step": 100
+    },
+    {
+      "epoch": 0.591715976331361,
+      "eval_loss": 0.703880250453949,
+      "eval_runtime": 20.1945,
+      "eval_samples_per_second": 14.856,
+      "eval_steps_per_second": 1.882,
+      "step": 100
+    },
+    {
+      "epoch": 0.650887573964497,
+      "grad_norm": 0.46801886344571736,
+      "learning_rate": 4.075219608954278e-05,
+      "loss": 0.6649,
+      "step": 110
+    },
+    {
+      "epoch": 0.7100591715976331,
+      "grad_norm": 0.7261378819214507,
+      "learning_rate": 3.876064386435646e-05,
+      "loss": 0.6485,
+      "step": 120
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": 0.6311105785325938,
+      "learning_rate": 3.663489828471953e-05,
+      "loss": 0.5978,
+      "step": 130
+    },
+    {
+      "epoch": 0.8284023668639053,
+      "grad_norm": 1.0566991845506082,
+      "learning_rate": 3.4395689552855955e-05,
+      "loss": 0.6504,
+      "step": 140
+    },
+    {
+      "epoch": 0.8875739644970414,
+      "grad_norm": 0.42416309958564125,
+      "learning_rate": 3.206485435998498e-05,
+      "loss": 0.6214,
+      "step": 150
+    },
+    {
+      "epoch": 0.9467455621301775,
+      "grad_norm": 0.42527966548399343,
+      "learning_rate": 2.9665122935613727e-05,
+      "loss": 0.6144,
+      "step": 160
+    },
+    {
+      "epoch": 1.0059171597633136,
+      "grad_norm": 0.3979436989011932,
+      "learning_rate": 2.7219897383073373e-05,
+      "loss": 0.6493,
+      "step": 170
+    },
+    {
+      "epoch": 1.0650887573964498,
+      "grad_norm": 0.24203747971705217,
+      "learning_rate": 2.475302346296336e-05,
+      "loss": 0.6474,
+      "step": 180
+    },
+    {
+      "epoch": 1.1242603550295858,
+      "grad_norm": 0.319431105160605,
+      "learning_rate": 2.2288558050064367e-05,
+      "loss": 0.564,
+      "step": 190
+    },
+    {
+      "epoch": 1.183431952662722,
+      "grad_norm": 0.5080338290607097,
+      "learning_rate": 1.9850534531472546e-05,
+      "loss": 0.5798,
+      "step": 200
+    },
+    {
+      "epoch": 1.183431952662722,
+      "eval_loss": 0.7010006904602051,
+      "eval_runtime": 18.8091,
+      "eval_samples_per_second": 15.95,
+      "eval_steps_per_second": 2.02,
+      "step": 200
+    },
+    {
+      "epoch": 1.242603550295858,
+      "grad_norm": 0.564376924960013,
+      "learning_rate": 1.746272843378493e-05,
+      "loss": 0.6323,
+      "step": 210
+    },
+    {
+      "epoch": 1.301775147928994,
+      "grad_norm": 0.498318513735455,
+      "learning_rate": 1.5148425564932084e-05,
+      "loss": 0.6106,
+      "step": 220
+    },
+    {
+      "epoch": 1.3609467455621302,
+      "grad_norm": 0.2827517725379645,
+      "learning_rate": 1.2930194931731382e-05,
+      "loss": 0.6374,
+      "step": 230
+    },
+    {
+      "epoch": 1.4201183431952662,
+      "grad_norm": 0.5094729039694312,
+      "learning_rate": 1.0829668647661559e-05,
+      "loss": 0.5758,
+      "step": 240
+    },
+    {
+      "epoch": 1.4792899408284024,
+      "grad_norm": 0.269479823527664,
+      "learning_rate": 8.867330977190877e-06,
+      "loss": 0.5988,
+      "step": 250
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": 0.3701018971236146,
+      "learning_rate": 7.062318573891716e-06,
+      "loss": 0.5956,
+      "step": 260
+    },
+    {
+      "epoch": 1.5976331360946747,
+      "grad_norm": 0.43248474882397636,
+      "learning_rate": 5.4322338604131715e-06,
+      "loss": 0.5467,
+      "step": 270
+    },
+    {
+      "epoch": 1.6568047337278107,
+      "grad_norm": 0.896259597910678,
+      "learning_rate": 3.992973370223896e-06,
+      "loss": 0.599,
+      "step": 280
+    },
+    {
+      "epoch": 1.7159763313609466,
+      "grad_norm": 0.28651524116933125,
+      "learning_rate": 2.75857272513132e-06,
+      "loss": 0.5505,
+      "step": 290
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "grad_norm": 0.5096606249052688,
+      "learning_rate": 1.7410697603511383e-06,
+      "loss": 0.5292,
+      "step": 300
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "eval_loss": 0.6997424960136414,
+      "eval_runtime": 19.0297,
+      "eval_samples_per_second": 15.765,
+      "eval_steps_per_second": 1.997,
+      "step": 300
+    },
+    {
+      "epoch": 1.834319526627219,
+      "grad_norm": 0.6033783324728439,
+      "learning_rate": 9.503871319271551e-07,
+      "loss": 0.6355,
+      "step": 310
+    },
+    {
+      "epoch": 1.893491124260355,
+      "grad_norm": 0.5681857650004716,
+      "learning_rate": 3.9423555131007925e-07,
+      "loss": 0.6309,
+      "step": 320
+    },
+    {
+      "epoch": 1.952662721893491,
+      "grad_norm": 0.5632696843447448,
+      "learning_rate": 7.803859074854425e-08,
+      "loss": 0.5814,
+      "step": 330
+    },
+    {
+      "epoch": 2.0,
+      "step": 338,
+      "total_flos": 123391664193536.0,
+      "train_loss": 0.6494796699320776,
+      "train_runtime": 1192.6701,
+      "train_samples_per_second": 4.528,
+      "train_steps_per_second": 0.283
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 338,
+  "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": 123391664193536.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen14B_models/qwen_14B_d0_iter1_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a25d2f7ff7b99ffc68ada4152b160a463ed0921fa81a00c7b3486de6a3598416
+size 7032

qwen14B_models/qwen_14B_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/models/Qwen1_5_14B
+model-index:
+- name: qwen_14B_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_14B_d1_iter2_model
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B) on the qwen_14B_d1_iter2_model dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.5391
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.5923        | 0.5917 | 100  | 0.5549          |
+| 0.485         | 1.1834 | 200  | 0.5391          |
+| 0.4405        | 1.7751 | 300  | 0.5399          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.41.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

qwen14B_models/qwen_14B_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_14B",
+  "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",
+    "v_proj",
+    "up_proj",
+    "q_proj",
+    "k_proj",
+    "down_proj",
+    "gate_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen14B_models/qwen_14B_d1_iter2_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df15f1ce5fdb1222217b5cbc3f7adb0972ffdd341003ec84982008441d4fd13c
+size 62415664

qwen14B_models/qwen_14B_d1_iter2_model/added_tokens.json

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

qwen14B_models/qwen_14B_d1_iter2_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.5390731692314148,
+    "eval_runtime": 18.8117,
+    "eval_samples_per_second": 15.948,
+    "eval_steps_per_second": 2.02,
+    "total_flos": 93650261901312.0,
+    "train_loss": 0.5521188734551153,
+    "train_runtime": 1213.4234,
+    "train_samples_per_second": 4.45,
+    "train_steps_per_second": 0.279
+}

qwen14B_models/qwen_14B_d1_iter2_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.5390731692314148,
+    "eval_runtime": 18.8117,
+    "eval_samples_per_second": 15.948,
+    "eval_steps_per_second": 2.02
+}

qwen14B_models/qwen_14B_d1_iter2_model/merge.log

@@ -0,0 +1,7 @@
+[2024-07-15 11:40:18,829] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+07/15/2024 11:40:20 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
+07/15/2024 11:40:20 - INFO - llmtuner.model.patcher - Using KV cache for faster generation.
+07/15/2024 11:40:22 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
+07/15/2024 11:40:34 - INFO - llmtuner.model.adapter - Merged 1 adapter(s).
+07/15/2024 11:40:34 - INFO - llmtuner.model.adapter - Loaded adapter(s): /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen14B_models/qwen_14B_d1_iter2_model
+07/15/2024 11:40:34 - INFO - llmtuner.model.loader - all params: 14167290880

qwen14B_models/qwen_14B_d1_iter2_model/runs/Jul15_11-17-33_t-20240715144206-zn57g-worker-0/events.out.tfevents.1721042359.t-20240715144206-zn57g-worker-0.113434.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16559cac3cfb7178647d5a8f32bc559009f69aaa43358bcb95a0cc562ed2334c
+size 13407

qwen14B_models/qwen_14B_d1_iter2_model/runs/Jul15_11-17-33_t-20240715144206-zn57g-worker-0/events.out.tfevents.1721043608.t-20240715144206-zn57g-worker-0.113434.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:133bf6149964f579e3296ac3bf3545238dc42b442f8e4742773c8a53855baa59
+size 359

qwen14B_models/qwen_14B_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
+  }
+}

qwen14B_models/qwen_14B_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
+}

qwen14B_models/qwen_14B_d1_iter2_model/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "total_flos": 93650261901312.0,
+    "train_loss": 0.5521188734551153,
+    "train_runtime": 1213.4234,
+    "train_samples_per_second": 4.45,
+    "train_steps_per_second": 0.279
+}

qwen14B_models/qwen_14B_d1_iter2_model/trainer_log.jsonl

@@ -0,0 +1,38 @@
+{"current_steps": 10, "total_steps": 338, "loss": 1.39, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.5e-05, "epoch": 0.05917159763313609, "percentage": 2.96, "elapsed_time": "0:00:40", "remaining_time": "0:22:17"}
+{"current_steps": 20, "total_steps": 338, "loss": 1.2536, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.11834319526627218, "percentage": 5.92, "elapsed_time": "0:01:13", "remaining_time": "0:19:21"}
+{"current_steps": 30, "total_steps": 338, "loss": 0.6182, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.98781004037916e-05, "epoch": 0.17751479289940827, "percentage": 8.88, "elapsed_time": "0:01:45", "remaining_time": "0:18:02"}
+{"current_steps": 40, "total_steps": 338, "loss": 0.5871, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.951359037609088e-05, "epoch": 0.23668639053254437, "percentage": 11.83, "elapsed_time": "0:02:17", "remaining_time": "0:17:05"}
+{"current_steps": 50, "total_steps": 338, "loss": 0.6622, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.891002460691306e-05, "epoch": 0.2958579881656805, "percentage": 14.79, "elapsed_time": "0:02:49", "remaining_time": "0:16:17"}
+{"current_steps": 60, "total_steps": 338, "loss": 0.5239, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.807328905014201e-05, "epoch": 0.35502958579881655, "percentage": 17.75, "elapsed_time": "0:03:22", "remaining_time": "0:15:36"}
+{"current_steps": 70, "total_steps": 338, "loss": 0.6031, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.7011543523897996e-05, "epoch": 0.41420118343195267, "percentage": 20.71, "elapsed_time": "0:03:54", "remaining_time": "0:14:59"}
+{"current_steps": 80, "total_steps": 338, "loss": 0.5342, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.573514213625505e-05, "epoch": 0.47337278106508873, "percentage": 23.67, "elapsed_time": "0:04:26", "remaining_time": "0:14:20"}
+{"current_steps": 90, "total_steps": 338, "loss": 0.5295, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.425653231231344e-05, "epoch": 0.5325443786982249, "percentage": 26.63, "elapsed_time": "0:04:58", "remaining_time": "0:13:43"}
+{"current_steps": 100, "total_steps": 338, "loss": 0.5923, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.259013340731224e-05, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:05:31", "remaining_time": "0:13:09"}
+{"current_steps": 100, "total_steps": 338, "loss": null, "eval_loss": 0.5548861622810364, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.591715976331361, "percentage": 29.59, "elapsed_time": "0:05:31", "remaining_time": "0:13:09"}
+{"current_steps": 110, "total_steps": 338, "loss": 0.5283, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.075219608954278e-05, "epoch": 0.650887573964497, "percentage": 32.54, "elapsed_time": "0:06:41", "remaining_time": "0:13:51"}
+{"current_steps": 120, "total_steps": 338, "loss": 0.5123, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.876064386435646e-05, "epoch": 0.7100591715976331, "percentage": 35.5, "elapsed_time": "0:07:13", "remaining_time": "0:13:06"}
+{"current_steps": 130, "total_steps": 338, "loss": 0.5148, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.663489828471953e-05, "epoch": 0.7692307692307693, "percentage": 38.46, "elapsed_time": "0:07:45", "remaining_time": "0:12:25"}
+{"current_steps": 140, "total_steps": 338, "loss": 0.5706, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.4395689552855955e-05, "epoch": 0.8284023668639053, "percentage": 41.42, "elapsed_time": "0:08:18", "remaining_time": "0:11:45"}
+{"current_steps": 150, "total_steps": 338, "loss": 0.5443, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.206485435998498e-05, "epoch": 0.8875739644970414, "percentage": 44.38, "elapsed_time": "0:08:51", "remaining_time": "0:11:05"}
+{"current_steps": 160, "total_steps": 338, "loss": 0.5345, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.9665122935613727e-05, "epoch": 0.9467455621301775, "percentage": 47.34, "elapsed_time": "0:09:23", "remaining_time": "0:10:27"}
+{"current_steps": 170, "total_steps": 338, "loss": 0.4759, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.7219897383073373e-05, "epoch": 1.0059171597633136, "percentage": 50.3, "elapsed_time": "0:09:57", "remaining_time": "0:09:50"}
+{"current_steps": 180, "total_steps": 338, "loss": 0.4207, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.475302346296336e-05, "epoch": 1.0650887573964498, "percentage": 53.25, "elapsed_time": "0:10:29", "remaining_time": "0:09:12"}
+{"current_steps": 190, "total_steps": 338, "loss": 0.4453, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.2288558050064367e-05, "epoch": 1.1242603550295858, "percentage": 56.21, "elapsed_time": "0:11:01", "remaining_time": "0:08:35"}
+{"current_steps": 200, "total_steps": 338, "loss": 0.485, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9850534531472546e-05, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:11:34", "remaining_time": "0:07:59"}
+{"current_steps": 200, "total_steps": 338, "loss": null, "eval_loss": 0.5390731692314148, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.183431952662722, "percentage": 59.17, "elapsed_time": "0:11:34", "remaining_time": "0:07:59"}
+{"current_steps": 210, "total_steps": 338, "loss": 0.448, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.746272843378493e-05, "epoch": 1.242603550295858, "percentage": 62.13, "elapsed_time": "0:12:42", "remaining_time": "0:07:44"}
+{"current_steps": 220, "total_steps": 338, "loss": 0.4257, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.5148425564932084e-05, "epoch": 1.301775147928994, "percentage": 65.09, "elapsed_time": "0:13:14", "remaining_time": "0:07:06"}
+{"current_steps": 230, "total_steps": 338, "loss": 0.5102, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.2930194931731382e-05, "epoch": 1.3609467455621302, "percentage": 68.05, "elapsed_time": "0:13:47", "remaining_time": "0:06:28"}
+{"current_steps": 240, "total_steps": 338, "loss": 0.4632, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.0829668647661559e-05, "epoch": 1.4201183431952662, "percentage": 71.01, "elapsed_time": "0:14:20", "remaining_time": "0:05:51"}
+{"current_steps": 250, "total_steps": 338, "loss": 0.5108, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.867330977190877e-06, "epoch": 1.4792899408284024, "percentage": 73.96, "elapsed_time": "0:14:52", "remaining_time": "0:05:14"}
+{"current_steps": 260, "total_steps": 338, "loss": 0.4304, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.062318573891716e-06, "epoch": 1.5384615384615383, "percentage": 76.92, "elapsed_time": "0:15:24", "remaining_time": "0:04:37"}
+{"current_steps": 270, "total_steps": 338, "loss": 0.4281, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.4322338604131715e-06, "epoch": 1.5976331360946747, "percentage": 79.88, "elapsed_time": "0:15:57", "remaining_time": "0:04:01"}
+{"current_steps": 280, "total_steps": 338, "loss": 0.4529, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.992973370223896e-06, "epoch": 1.6568047337278107, "percentage": 82.84, "elapsed_time": "0:16:29", "remaining_time": "0:03:24"}
+{"current_steps": 290, "total_steps": 338, "loss": 0.4698, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.75857272513132e-06, "epoch": 1.7159763313609466, "percentage": 85.8, "elapsed_time": "0:17:01", "remaining_time": "0:02:49"}
+{"current_steps": 300, "total_steps": 338, "loss": 0.4405, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.7410697603511383e-06, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:17:34", "remaining_time": "0:02:13"}
+{"current_steps": 300, "total_steps": 338, "loss": null, "eval_loss": 0.5398803353309631, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.7751479289940828, "percentage": 88.76, "elapsed_time": "0:17:34", "remaining_time": "0:02:13"}
+{"current_steps": 310, "total_steps": 338, "loss": 0.4825, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.503871319271551e-07, "epoch": 1.834319526627219, "percentage": 91.72, "elapsed_time": "0:18:40", "remaining_time": "0:01:41"}
+{"current_steps": 320, "total_steps": 338, "loss": 0.5166, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.9423555131007925e-07, "epoch": 1.893491124260355, "percentage": 94.67, "elapsed_time": "0:19:12", "remaining_time": "0:01:04"}
+{"current_steps": 330, "total_steps": 338, "loss": 0.4212, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.803859074854425e-08, "epoch": 1.952662721893491, "percentage": 97.63, "elapsed_time": "0:19:45", "remaining_time": "0:00:28"}
+{"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": "0:20:13", "remaining_time": "0:00:00"}
+{"current_steps": 38, "total_steps": 38, "loss": null, "eval_loss": 0.5390731692314148, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.0, "percentage": 100.0, "elapsed_time": "0:20:48", "remaining_time": "0:00:00"}

qwen14B_models/qwen_14B_d1_iter2_model/trainer_state.json

@@ -0,0 +1,297 @@
+{
+  "best_metric": 0.5390731692314148,
+  "best_model_checkpoint": "/ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen14B_models/qwen_14B_d1_iter2_model/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": 1.5490762637870956,
+      "learning_rate": 2.5e-05,
+      "loss": 1.39,
+      "step": 10
+    },
+    {
+      "epoch": 0.11834319526627218,
+      "grad_norm": 2.180427642573537,
+      "learning_rate": 5e-05,
+      "loss": 1.2536,
+      "step": 20
+    },
+    {
+      "epoch": 0.17751479289940827,
+      "grad_norm": 1.4938352528249466,
+      "learning_rate": 4.98781004037916e-05,
+      "loss": 0.6182,
+      "step": 30
+    },
+    {
+      "epoch": 0.23668639053254437,
+      "grad_norm": 0.8266774666049563,
+      "learning_rate": 4.951359037609088e-05,
+      "loss": 0.5871,
+      "step": 40
+    },
+    {
+      "epoch": 0.2958579881656805,
+      "grad_norm": 0.44331898296236955,
+      "learning_rate": 4.891002460691306e-05,
+      "loss": 0.6622,
+      "step": 50
+    },
+    {
+      "epoch": 0.35502958579881655,
+      "grad_norm": 0.7348699943282603,
+      "learning_rate": 4.807328905014201e-05,
+      "loss": 0.5239,
+      "step": 60
+    },
+    {
+      "epoch": 0.41420118343195267,
+      "grad_norm": 1.0233758530765529,
+      "learning_rate": 4.7011543523897996e-05,
+      "loss": 0.6031,
+      "step": 70
+    },
+    {
+      "epoch": 0.47337278106508873,
+      "grad_norm": 0.44566662901156306,
+      "learning_rate": 4.573514213625505e-05,
+      "loss": 0.5342,
+      "step": 80
+    },
+    {
+      "epoch": 0.5325443786982249,
+      "grad_norm": 0.32376173126197005,
+      "learning_rate": 4.425653231231344e-05,
+      "loss": 0.5295,
+      "step": 90
+    },
+    {
+      "epoch": 0.591715976331361,
+      "grad_norm": 2.5050276715521953,
+      "learning_rate": 4.259013340731224e-05,
+      "loss": 0.5923,
+      "step": 100
+    },
+    {
+      "epoch": 0.591715976331361,
+      "eval_loss": 0.5548861622810364,
+      "eval_runtime": 20.5369,
+      "eval_samples_per_second": 14.608,
+      "eval_steps_per_second": 1.85,
+      "step": 100
+    },
+    {
+      "epoch": 0.650887573964497,
+      "grad_norm": 0.6939641269685894,
+      "learning_rate": 4.075219608954278e-05,
+      "loss": 0.5283,
+      "step": 110
+    },
+    {
+      "epoch": 0.7100591715976331,
+      "grad_norm": 0.3223403825561066,
+      "learning_rate": 3.876064386435646e-05,
+      "loss": 0.5123,
+      "step": 120
+    },
+    {
+      "epoch": 0.7692307692307693,
+      "grad_norm": 1.242304136081819,
+      "learning_rate": 3.663489828471953e-05,
+      "loss": 0.5148,
+      "step": 130
+    },
+    {
+      "epoch": 0.8284023668639053,
+      "grad_norm": 0.6162370328893566,
+      "learning_rate": 3.4395689552855955e-05,
+      "loss": 0.5706,
+      "step": 140
+    },
+    {
+      "epoch": 0.8875739644970414,
+      "grad_norm": 2.854595963693751,
+      "learning_rate": 3.206485435998498e-05,
+      "loss": 0.5443,
+      "step": 150
+    },
+    {
+      "epoch": 0.9467455621301775,
+      "grad_norm": 0.836450705626818,
+      "learning_rate": 2.9665122935613727e-05,
+      "loss": 0.5345,
+      "step": 160
+    },
+    {
+      "epoch": 1.0059171597633136,
+      "grad_norm": 0.6854327873281724,
+      "learning_rate": 2.7219897383073373e-05,
+      "loss": 0.4759,
+      "step": 170
+    },
+    {
+      "epoch": 1.0650887573964498,
+      "grad_norm": 0.3289222105005388,
+      "learning_rate": 2.475302346296336e-05,
+      "loss": 0.4207,
+      "step": 180
+    },
+    {
+      "epoch": 1.1242603550295858,
+      "grad_norm": 0.6402929677890956,
+      "learning_rate": 2.2288558050064367e-05,
+      "loss": 0.4453,
+      "step": 190
+    },
+    {
+      "epoch": 1.183431952662722,
+      "grad_norm": 0.4560254960950853,
+      "learning_rate": 1.9850534531472546e-05,
+      "loss": 0.485,
+      "step": 200
+    },
+    {
+      "epoch": 1.183431952662722,
+      "eval_loss": 0.5390731692314148,
+      "eval_runtime": 19.171,
+      "eval_samples_per_second": 15.649,
+      "eval_steps_per_second": 1.982,
+      "step": 200
+    },
+    {
+      "epoch": 1.242603550295858,
+      "grad_norm": 0.4429967217295556,
+      "learning_rate": 1.746272843378493e-05,
+      "loss": 0.448,
+      "step": 210
+    },
+    {
+      "epoch": 1.301775147928994,
+      "grad_norm": 0.5801927088218751,
+      "learning_rate": 1.5148425564932084e-05,
+      "loss": 0.4257,
+      "step": 220
+    },
+    {
+      "epoch": 1.3609467455621302,
+      "grad_norm": 1.6461621420887593,
+      "learning_rate": 1.2930194931731382e-05,
+      "loss": 0.5102,
+      "step": 230
+    },
+    {
+      "epoch": 1.4201183431952662,
+      "grad_norm": 1.4629083533873057,
+      "learning_rate": 1.0829668647661559e-05,
+      "loss": 0.4632,
+      "step": 240
+    },
+    {
+      "epoch": 1.4792899408284024,
+      "grad_norm": 1.6289265138854685,
+      "learning_rate": 8.867330977190877e-06,
+      "loss": 0.5108,
+      "step": 250
+    },
+    {
+      "epoch": 1.5384615384615383,
+      "grad_norm": 0.6682669143854097,
+      "learning_rate": 7.062318573891716e-06,
+      "loss": 0.4304,
+      "step": 260
+    },
+    {
+      "epoch": 1.5976331360946747,
+      "grad_norm": 0.5194393122476173,
+      "learning_rate": 5.4322338604131715e-06,
+      "loss": 0.4281,
+      "step": 270
+    },
+    {
+      "epoch": 1.6568047337278107,
+      "grad_norm": 1.0124557614072613,
+      "learning_rate": 3.992973370223896e-06,
+      "loss": 0.4529,
+      "step": 280
+    },
+    {
+      "epoch": 1.7159763313609466,
+      "grad_norm": 0.42873082380830013,
+      "learning_rate": 2.75857272513132e-06,
+      "loss": 0.4698,
+      "step": 290
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "grad_norm": 0.9876326963962394,
+      "learning_rate": 1.7410697603511383e-06,
+      "loss": 0.4405,
+      "step": 300
+    },
+    {
+      "epoch": 1.7751479289940828,
+      "eval_loss": 0.5398803353309631,
+      "eval_runtime": 18.7716,
+      "eval_samples_per_second": 15.982,
+      "eval_steps_per_second": 2.024,
+      "step": 300
+    },
+    {
+      "epoch": 1.834319526627219,
+      "grad_norm": 0.9194473587313593,
+      "learning_rate": 9.503871319271551e-07,
+      "loss": 0.4825,
+      "step": 310
+    },
+    {
+      "epoch": 1.893491124260355,
+      "grad_norm": 1.330707485363624,
+      "learning_rate": 3.9423555131007925e-07,
+      "loss": 0.5166,
+      "step": 320
+    },
+    {
+      "epoch": 1.952662721893491,
+      "grad_norm": 0.3585144579596122,
+      "learning_rate": 7.803859074854425e-08,
+      "loss": 0.4212,
+      "step": 330
+    },
+    {
+      "epoch": 2.0,
+      "step": 338,
+      "total_flos": 93650261901312.0,
+      "train_loss": 0.5521188734551153,
+      "train_runtime": 1213.4234,
+      "train_samples_per_second": 4.45,
+      "train_steps_per_second": 0.279
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 338,
+  "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": 93650261901312.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

qwen14B_models/qwen_14B_d1_iter2_model/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0761f6bb67160b2669ef61d48e036499824b05c6c2a1509429473af05407f0ea
+size 7032

qwen14B_models/qwen_14B_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/models/Qwen1_5_14B
+model-index:
+- name: qwen_14B_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_14B_d2_iter3_model
+
+This model is a fine-tuned version of [/ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B](https://huggingface.co//ML-A100/team/mm/eamon/self_instruction/models/Qwen1_5_14B) on the qwen_14B_d2_iter3_model dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.5153
+
+## 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 |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.4759        | 0.5917 | 100  | 0.5316          |
+| 0.3967        | 1.1834 | 200  | 0.5153          |
+| 0.3547        | 1.7751 | 300  | 0.5178          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.41.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.18.0
+- Tokenizers 0.19.1

qwen14B_models/qwen_14B_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_14B",
+  "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",
+    "gate_proj",
+    "up_proj",
+    "down_proj",
+    "k_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

qwen14B_models/qwen_14B_d2_iter3_model/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9dd4a13be41dcad171d6b6ee79720860d70b46c6f874c8b2ece3ad9084c994e7
+size 62415664

qwen14B_models/qwen_14B_d2_iter3_model/added_tokens.json

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

qwen14B_models/qwen_14B_d2_iter3_model/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.5153154134750366,
+    "eval_runtime": 19.1336,
+    "eval_samples_per_second": 15.679,
+    "eval_steps_per_second": 1.986,
+    "total_flos": 84346414039040.0,
+    "train_loss": 0.5209799306632499,
+    "train_runtime": 1210.5031,
+    "train_samples_per_second": 4.461,
+    "train_steps_per_second": 0.279
+}

qwen14B_models/qwen_14B_d2_iter3_model/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 2.0,
+    "eval_loss": 0.5153154134750366,
+    "eval_runtime": 19.1336,
+    "eval_samples_per_second": 15.679,
+    "eval_steps_per_second": 1.986
+}

qwen14B_models/qwen_14B_d2_iter3_model/merge.log

@@ -0,0 +1,7 @@
+[2024-07-15 14:24:23,419] [INFO] [real_accelerator.py:191:get_accelerator] Setting ds_accelerator to cuda (auto detect)
+07/15/2024 14:24:25 - INFO - llmtuner.data.template - Add <|im_end|>,<|endoftext|> to stop words.
+07/15/2024 14:24:25 - INFO - llmtuner.model.patcher - Using KV cache for faster generation.
+07/15/2024 14:24:27 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
+07/15/2024 14:24:38 - INFO - llmtuner.model.adapter - Merged 1 adapter(s).
+07/15/2024 14:24:38 - INFO - llmtuner.model.adapter - Loaded adapter(s): /ML-A100/team/mm/eamon/self_instruction/seed_ppl/qwen14B_models/qwen_14B_d2_iter3_model
+07/15/2024 14:24:38 - INFO - llmtuner.model.loader - all params: 14167290880

qwen14B_models/qwen_14B_d2_iter3_model/runs/Jul15_14-01-40_t-20240715144206-zn57g-worker-0/events.out.tfevents.1721052207.t-20240715144206-zn57g-worker-0.174136.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bd260b071146146fb2f9a5447ff861d3c26a732d2b53000c7fd530fda28d9bf
+size 13407