add config.json
Browse files- config.json +51 -0
config.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"seed": 0,
|
| 3 |
+
"exp_name": "pythia_SFT_ga4",
|
| 4 |
+
"batch_size": 64,
|
| 5 |
+
"eval_batch_size": 32,
|
| 6 |
+
"debug": false,
|
| 7 |
+
"fsdp_port": 12355,
|
| 8 |
+
"datasets": [
|
| 9 |
+
"hh"
|
| 10 |
+
],
|
| 11 |
+
"wandb": {
|
| 12 |
+
"enabled": true,
|
| 13 |
+
"entity": "pythia_dpo",
|
| 14 |
+
"project": "Pythia_LOM"
|
| 15 |
+
},
|
| 16 |
+
"local_dirs": [
|
| 17 |
+
"/scr-ssd",
|
| 18 |
+
"/scr",
|
| 19 |
+
".cache"
|
| 20 |
+
],
|
| 21 |
+
"sample_during_eval": false,
|
| 22 |
+
"n_eval_model_samples": 16,
|
| 23 |
+
"do_first_eval": true,
|
| 24 |
+
"local_run_dir": ".cache/laura/pythia_SFT_ga4_2023-07-16_16-50-13_244945",
|
| 25 |
+
"lr": 0.000001,
|
| 26 |
+
"gradient_accumulation_steps": 4,
|
| 27 |
+
"max_grad_norm": 10,
|
| 28 |
+
"max_length": 512,
|
| 29 |
+
"max_prompt_length": 256,
|
| 30 |
+
"n_epochs": 1,
|
| 31 |
+
"n_examples": null,
|
| 32 |
+
"n_eval_examples": 256,
|
| 33 |
+
"trainer": "FSDPTrainer",
|
| 34 |
+
"optimizer": "RMSprop",
|
| 35 |
+
"warmup_steps": 150,
|
| 36 |
+
"activation_checkpointing": false,
|
| 37 |
+
"eval_every": 19968,
|
| 38 |
+
"minimum_log_interval_secs": 1,
|
| 39 |
+
"model": {
|
| 40 |
+
"name_or_path": "EleutherAI/pythia-70m",
|
| 41 |
+
"tokenizer_name_or_path": null,
|
| 42 |
+
"archive": null,
|
| 43 |
+
"block_name": "GPTNeoXLayer",
|
| 44 |
+
"policy_dtype": "float32",
|
| 45 |
+
"fsdp_policy_mp": "bfloat16",
|
| 46 |
+
"reference_dtype": "float16"
|
| 47 |
+
},
|
| 48 |
+
"loss": {
|
| 49 |
+
"name": "sft"
|
| 50 |
+
}
|
| 51 |
+
}
|