File size: 3,521 Bytes
5bcec5b f97c6d8 5bcec5b f12c5db 5bcec5b 4061222 5bcec5b 4c6deb2 5bcec5b 64ad91d 5bcec5b e8ab4f5 5bcec5b 4c6deb2 5bcec5b 4c6deb2 f97c6d8 4c6deb2 f97c6d8 4c6deb2 f97c6d8 5bcec5b 4c6deb2 5bcec5b f97c6d8 5bcec5b 4c6deb2 5bcec5b 4c6deb2 5bcec5b 4c6deb2 5bcec5b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 | {
"train_loop_config": {
"a2g_norm": false,
"base_path": "/common/home/users/m/mq.nguyen.2023/testcode/SAC_LM/module9_clmv2",
"beta_coef": "0.0",
"block_size": 1024,
"buffer_max_size": 20000,
"causal_model": true,
"checkpoint_interval": "100",
"clamp_update": false,
"config_name": null,
"constraint_type": "kl",
"cpu_per_worker": 7,
"dataset_config_name": null,
"dataset_name": "nbtpj/summ_ds_train",
"ent_coef": 0.0001,
"epsilon": 0.2,
"eval_interval": "100",
"freeze_role2": false,
"gamma": 0.95,
"gpu_per_worker": 1,
"grad_norm": 0.5,
"gradient_accumulation_steps": 1,
"gradient_checkpoint": false,
"group_relative_norm": false,
"inference_config": {
"do_sample": true,
"max_new_tokens": 135,
"min_new_tokens": 32,
"temperature": 0.0
},
"keep_eval_size": false,
"label_col": "summary",
"learning_rate": 3e-07,
"lm_fraction": -1.0,
"lm_w": 0.0,
"log_interval": "5m",
"log_rollout_txt": true,
"lora": true,
"lr_scheduler_type": "constant",
"max_train_rollouts": 100000,
"max_train_steps": 40000,
"mini_epoch": 1,
"mixed_precision": "bf16",
"model_name_or_path": "gpt2",
"model_type": null,
"n_augment": 0,
"n_generate": 4,
"need_attn_mask": true,
"num_warmup_steps": 200,
"only_train_role1": false,
"per_device_eval_batch_size": 4,
"per_device_query_rollout_batch_size": 32,
"per_device_train_batch_size": 4,
"pretrained_role2_name_or_path": "none",
"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:",
"push_to_hub": null,
"ref_role1_name_or_path": "gpt2",
"ref_role2_name_or_path": "gpt2",
"rl_algo": "off_policy",
"rl_w": 1.0,
"rollout_config": {
"accuracy_w": 1.96282248134428,
"accuracy_w2": 19.549524484259873,
"len_pen": 1.0,
"len_pen2": 1.0,
"max_ctx_len": 860,
"sampling_params_1": {
"logprobs": 5,
"max_tokens": 135,
"min_tokens": 32,
"n": 1,
"temperature": 1.0
},
"sampling_params_2": {
"logprobs": 5,
"max_tokens": 860,
"min_tokens": 5,
"n": 4,
"temperature": 1.0
},
"similarity_fn": "rouge",
"threshold": 0.01605191133358762
},
"rollout_game": "baseline3v2",
"sample_config": {
"do_sample": true,
"min_new_tokens": 1,
"temperature": 1.0
},
"script": "/common/home/users/m/mq.nguyen.2023/testcode/SAC_LM/module9_clmv2/execute/utune/debug.py",
"seed": 0,
"test_clm": false,
"test_gen": true,
"test_glue": false,
"text_template": "{text}\nTL;DR: {summary}",
"tokenizer_name": null,
"train_from_raw": true,
"train_split_name": "sim_with_one_golden__cnndm_train",
"trunc_eval": 256,
"trunc_evals": [
"cnndm___10"
],
"trust_remote_code": true,
"tune_metrics": [
"cnndm/rouge1___1.0",
"cnndm/rouge2___2.0",
"cnndm/bertscore_f1___0.25"
],
"use_deepspeed": false,
"use_slow_tokenizer": false,
"vectorizer_path": "/common/home/users/m/mq.nguyen.2023/testcode/SAC_LM/module9_clmv2/vectorizer/wikitext103_tfidf_full.joblib",
"vllm_sleep": true,
"vllm_vram_ratio": 0.3,
"weight_decay": 1e-05,
"world_size": 1,
"zero_config": 2
}
} |