Upload best model checkpoint

accelerator.ckpt/dl_state_dict.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d2567898e7971a0b5670bf6a2737bc31f98cc950fb4130fe8d746054d9f1be3
+size 1489

accelerator.ckpt/dl_state_dict_1.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43919dbdf599450386720f61a25bad24d90eca951c25271e500df587378f1abf
+size 1501

accelerator.ckpt/dl_state_dict_2.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d18bcd689dbb471d486cfc1fcae82e7fb210ffb37bd6172295a630d9fea40b9c
+size 1501

accelerator.ckpt/dl_state_dict_3.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7071a9b7092627c45be3df7796c8387732b699abc24c431c8d8abe771d2f1cdb
+size 1501

accelerator.ckpt/dl_state_dict_4.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c852d5f1a0b46f769dceae5ed0350bbebda4ae59f4d937518014c7ffd8c39db
+size 1501

accelerator.ckpt/dl_state_dict_5.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e368128de79e58e8060f0c777fb7be2bf456de524f4e7a86faf5ccaf21d3d3e9
+size 1501

accelerator.ckpt/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72cb730d48a4ebb7c171131201d4ef436a363e29f3612a1bef070eb443d27540
+size 497774208

accelerator.ckpt/optimizer.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4bf5be8ddaa821b05c94549412b54654e98f21c54385c65705aaa6a10cbd0ea6
+size 995606091

accelerator.ckpt/random_states_0.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c2096e5fe9ed800834e5062b80cf17dbcc2e3a970c7f22cab692e6840cd078a
+size 14757

accelerator.ckpt/scheduler.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca27db8230fb5ac7b02317d1adc87bf3ba1bc5dd9063fe15c2d4bbad81e79932
+size 1465

config.json

@@ -0,0 +1,38 @@
+{
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 50256,
+  "dtype": "float32",
+  "embd_pdrop": 0.1,
+  "eos_token_id": 50256,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 1024,
+  "n_embd": 768,
+  "n_head": 12,
+  "n_inner": null,
+  "n_layer": 12,
+  "n_positions": 1024,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "task_specific_params": {
+    "text-generation": {
+      "do_sample": true,
+      "max_length": 50
+    }
+  },
+  "transformers_version": "4.57.3",
+  "use_cache": true,
+  "vocab_size": 50257
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "transformers_version": "4.57.3"
+}

metrics.json

@@ -0,0 +1,51 @@
+{
+    "rl_info/A2G": -2.228810262749903e-05,
+    "rl_info/entropy": 3.103461980819702,
+    "rl_info/total_token": 3872.0,
+    "rl_info/advantage_b4_norm": -847.2278442382812,
+    "rl_info/advantage_after_gnorm": 0.2974321246147156,
+    "rl_info/kl_w_ref": 0.0,
+    "train/rl_loss": 0.0019184639677405357,
+    "train/total_loss": 0.0019184639677405357,
+    "gigaword/rouge1": 0.017424667842673495,
+    "gigaword/rouge2": 0.002778778179571005,
+    "gigaword/rougeL": 0.01708351188579019,
+    "gigaword/rougeLsum": 0.01597830557868938,
+    "gigaword/bertscore_precision": 0.5688383205235005,
+    "gigaword/bertscore_recall": 0.6489311541616917,
+    "gigaword/bertscore_f1": 0.6054284919798374,
+    "cnndm/rouge1": 0.1657242795787633,
+    "cnndm/rouge2": 0.05944159953280759,
+    "cnndm/rougeL": 0.13357654072383415,
+    "cnndm/rougeLsum": 0.14685147194336262,
+    "cnndm/bertscore_precision": 0.6781402329603831,
+    "cnndm/bertscore_recall": 0.7525439510742823,
+    "cnndm/bertscore_f1": 0.7120969245831171,
+    "xsum/rouge1": 0.12156545540546186,
+    "xsum/rouge2": 0.01235304893841479,
+    "xsum/rougeL": 0.08560314826140131,
+    "xsum/rougeLsum": 0.09009432870268685,
+    "xsum/bertscore_precision": 0.6957823038101196,
+    "xsum/bertscore_recall": 0.7223540594180425,
+    "xsum/bertscore_f1": 0.7064366390307745,
+    "samsum/rouge1": 0.08285230780746375,
+    "samsum/rouge2": 0.021593044513516627,
+    "samsum/rougeL": 0.07018645398318014,
+    "samsum/rougeLsum": 0.057873523121751924,
+    "samsum/bertscore_precision": 0.6268573751052221,
+    "samsum/bertscore_recall": 0.6982430865367254,
+    "samsum/bertscore_f1": 0.6594596952199936,
+    "eval_agg/avg_all_rougef": 0.06881127912496056,
+    "eval_agg/avg_all_bertf": 0.6708554377034307,
+    "eval_agg/avg_all": 0.3698333584141956,
+    "num_rl_rollout": 10,
+    "lm_epoch": 0,
+    "rl_epoch": 0,
+    "step": 400,
+    "total_data_token": 109047,
+    "total_rl_token": 1767730,
+    "total_lm_token": 0,
+    "total_token": 1767730,
+    "completed_steps": 400,
+    "tune_objective": 0.8896403575001602
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72cb730d48a4ebb7c171131201d4ef436a363e29f3612a1bef070eb443d27540
+size 497774208

special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,21 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "50256": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": true,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 1024,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

train_configs.json

@@ -0,0 +1,110 @@
+{
+    "dataset_name": "nbtpj/summ_ds_train",
+    "dataset_config_name": null,
+    "train_split_name": "merge36_cnndmsamsumxsum",
+    "text_col": "text",
+    "freeze_role2": false,
+    "only_train_role1": false,
+    "model_name_or_path": "gpt2",
+    "ref_role1_name_or_path": "gpt2",
+    "ref_role2_name_or_path": "gpt2",
+    "pretrained_role2_name_or_path": "none",
+    "config_name": null,
+    "vectorizer_path": "/common/home/users/m/mq.nguyen.2023/testcode/SAC_LM/module9_clmv2/vectorizer/wikitext103_tfidf_full.joblib",
+    "tokenizer_name": null,
+    "use_slow_tokenizer": false,
+    "per_device_train_batch_size": 4,
+    "per_device_query_rollout_batch_size": 32,
+    "per_device_eval_batch_size": 4,
+    "vllm_vram_ratio": 0.3,
+    "learning_rate": 3e-07,
+    "grad_norm": 0.5,
+    "weight_decay": 1e-05,
+    "max_train_steps": 40000,
+    "max_train_rollouts": 100000,
+    "gradient_accumulation_steps": 1,
+    "lr_scheduler_type": "constant",
+    "num_warmup_steps": 200,
+    "seed": 0,
+    "model_type": null,
+    "block_size": 1024,
+    "mini_epoch": 1,
+    "rollout_game": "baseline3v2",
+    "rl_algo": "on_policy",
+    "constraint_type": "kl",
+    "clamp_update": false,
+    "rl_w": 1.0,
+    "lm_w": 0.0,
+    "n_generate": 4,
+    "n_augment": 0,
+    "gradient_checkpoint": false,
+    "group_relative_norm": true,
+    "sample_config": {
+        "do_sample": true,
+        "min_new_tokens": 20,
+        "temperature": 1.0
+    },
+    "inference_config": {
+        "do_sample": true,
+        "temperature": 0.0,
+        "min_new_tokens": 5,
+        "max_new_tokens": 250
+    },
+    "rollout_config": {
+        "accuracy_w": 60.07249475906205,
+        "len_pen": 1.0,
+        "accuracy_w2": 1.2229065947034368,
+        "len_pen2": 1.0,
+        "threshold": 0.006750312521595928,
+        "similarity_fn": "rouge"
+    },
+    "ent_coef": 0.0001,
+    "beta_coef": "0.0",
+    "prompt_0": "{text}",
+    "prompt_1": "{text}\nTL;DR: ",
+    "prompt_2": "Given the text: {role1_output}\nReconstruct the summarized text to the detailed:",
+    "prompt_eval": "{text}\nTL;DR:",
+    "epsilon": 0.2,
+    "a2g_norm": true,
+    "vllm_sleep": true,
+    "lora": false,
+    "need_attn_mask": true,
+    "gamma": 0.95,
+    "trust_remote_code": true,
+    "test_glue": false,
+    "test_clm": false,
+    "causal_model": true,
+    "test_gen": true,
+    "log_rollout_txt": true,
+    "trunc_eval": 256,
+    "trunc_evals": [
+        "cnndm___12",
+        "samsum___12",
+        "xsum___12",
+        "gigaword___200",
+        "duc___50"
+    ],
+    "use_deepspeed": false,
+    "zero_config": 2,
+    "log_interval": "5m",
+    "eval_interval": "100",
+    "checkpoint_interval": "100",
+    "lm_fraction": -1.0,
+    "push_to_hub": null,
+    "keep_eval_size": false,
+    "mixed_precision": "bf16",
+    "tune_metrics": [
+        "cnndm/rouge1___1.0",
+        "cnndm/bertscore_f1___0.25",
+        "samsum/rouge1___1.0",
+        "samsum/bertscore_f1___0.25",
+        "xsum/rouge1___1.0",
+        "xsum/bertscore_f1___0.25"
+    ],
+    "base_path": "/common/home/users/m/mq.nguyen.2023/testcode/SAC_LM/module9_clmv2",
+    "script": "/common/home/users/m/mq.nguyen.2023/testcode/SAC_LM/module9_clmv2/execute/utune/debug.py",
+    "train_from_raw": true,
+    "world_size": 1,
+    "cpu_per_worker": 7,
+    "gpu_per_worker": 1
+}