208 lines
3.6 KiB
JSON
208 lines
3.6 KiB
JSON
{
|
|
"mode": "rand",
|
|
"seed": 42,
|
|
"hf_repo": "Likithp/v9_rand_s42",
|
|
"base_model": "Qwen/Qwen2.5-0.5B-Instruct",
|
|
"dataset": "data/cs9_rand_v1",
|
|
"hf_dataset": "Likithp/cs9_rand_v1",
|
|
"trained_at": "2026-06-04T10:38:14.461918+00:00",
|
|
"optimizer": "AdamW",
|
|
"lr": 3e-05,
|
|
"weight_decay": 0.0,
|
|
"batch_size": 8,
|
|
"grad_accum": 8,
|
|
"effective_batch": 64,
|
|
"epochs": 5,
|
|
"warmup_steps": 200,
|
|
"grad_clip": 0.5,
|
|
"lr_schedule": "cosine",
|
|
"checkpoint_criterion": "train_loss",
|
|
"max_seq_len": 256,
|
|
"dtype": "bfloat16",
|
|
"alias_groups": {
|
|
"T1": [
|
|
"bev",
|
|
"cif",
|
|
"dov",
|
|
"fal",
|
|
"gev",
|
|
"hac",
|
|
"jac",
|
|
"kab",
|
|
"lex",
|
|
"nad"
|
|
],
|
|
"T2": [
|
|
"agov",
|
|
"egiv",
|
|
"ejof",
|
|
"ikob",
|
|
"okiv",
|
|
"uliv",
|
|
"vmob",
|
|
"vnob",
|
|
"xnob",
|
|
"znob"
|
|
]
|
|
},
|
|
"log_every_steps": 50,
|
|
"eval_n": 500,
|
|
"dry_run": false,
|
|
"best_epoch": 5,
|
|
"best_train_loss": 0.091564,
|
|
"val_exact_match": 0.048,
|
|
"val_outer_alias_acc": 1.0,
|
|
"val_inner_alias_acc": 0.057,
|
|
"val_t1_outer_alias_acc": 1.0,
|
|
"val_t2_outer_alias_acc": 1.0,
|
|
"val_t1_t2_gap_pp": 0.0,
|
|
"val_inner_by_alias": {
|
|
"agov": {
|
|
"correct": 3,
|
|
"total": 42,
|
|
"pct": 7.1,
|
|
"token_group": "T2"
|
|
},
|
|
"bev": {
|
|
"correct": 3,
|
|
"total": 53,
|
|
"pct": 5.7,
|
|
"token_group": "T1"
|
|
},
|
|
"cif": {
|
|
"correct": 2,
|
|
"total": 58,
|
|
"pct": 3.4,
|
|
"token_group": "T1"
|
|
},
|
|
"dov": {
|
|
"correct": 0,
|
|
"total": 39,
|
|
"pct": 0.0,
|
|
"token_group": "T1"
|
|
},
|
|
"egiv": {
|
|
"correct": 3,
|
|
"total": 49,
|
|
"pct": 6.1,
|
|
"token_group": "T2"
|
|
},
|
|
"ejof": {
|
|
"correct": 2,
|
|
"total": 42,
|
|
"pct": 4.8,
|
|
"token_group": "T2"
|
|
},
|
|
"fal": {
|
|
"correct": 4,
|
|
"total": 43,
|
|
"pct": 9.3,
|
|
"token_group": "T1"
|
|
},
|
|
"gev": {
|
|
"correct": 0,
|
|
"total": 48,
|
|
"pct": 0.0,
|
|
"token_group": "T1"
|
|
},
|
|
"hac": {
|
|
"correct": 1,
|
|
"total": 56,
|
|
"pct": 1.8,
|
|
"token_group": "T1"
|
|
},
|
|
"ikob": {
|
|
"correct": 1,
|
|
"total": 50,
|
|
"pct": 2.0,
|
|
"token_group": "T2"
|
|
},
|
|
"jac": {
|
|
"correct": 5,
|
|
"total": 51,
|
|
"pct": 9.8,
|
|
"token_group": "T1"
|
|
},
|
|
"kab": {
|
|
"correct": 1,
|
|
"total": 59,
|
|
"pct": 1.7,
|
|
"token_group": "T1"
|
|
},
|
|
"lex": {
|
|
"correct": 0,
|
|
"total": 51,
|
|
"pct": 0.0,
|
|
"token_group": "T1"
|
|
},
|
|
"nad": {
|
|
"correct": 5,
|
|
"total": 64,
|
|
"pct": 7.8,
|
|
"token_group": "T1"
|
|
},
|
|
"okiv": {
|
|
"correct": 1,
|
|
"total": 42,
|
|
"pct": 2.4,
|
|
"token_group": "T2"
|
|
},
|
|
"uliv": {
|
|
"correct": 7,
|
|
"total": 46,
|
|
"pct": 15.2,
|
|
"token_group": "T2"
|
|
},
|
|
"vmob": {
|
|
"correct": 11,
|
|
"total": 55,
|
|
"pct": 20.0,
|
|
"token_group": "T2"
|
|
},
|
|
"vnob": {
|
|
"correct": 5,
|
|
"total": 48,
|
|
"pct": 10.4,
|
|
"token_group": "T2"
|
|
},
|
|
"xnob": {
|
|
"correct": 2,
|
|
"total": 41,
|
|
"pct": 4.9,
|
|
"token_group": "T2"
|
|
},
|
|
"znob": {
|
|
"correct": 1,
|
|
"total": 63,
|
|
"pct": 1.6,
|
|
"token_group": "T2"
|
|
}
|
|
},
|
|
"train_log": [
|
|
{
|
|
"epoch": 1,
|
|
"loss": 0.296072,
|
|
"val_em": 0.05
|
|
},
|
|
{
|
|
"epoch": 2,
|
|
"loss": 0.095528,
|
|
"val_em": 0.054
|
|
},
|
|
{
|
|
"epoch": 3,
|
|
"loss": 0.094019,
|
|
"val_em": 0.046
|
|
},
|
|
{
|
|
"epoch": 4,
|
|
"loss": 0.092476,
|
|
"val_em": 0.058
|
|
},
|
|
{
|
|
"epoch": 5,
|
|
"loss": 0.091564,
|
|
"val_em": 0.048
|
|
}
|
|
]
|
|
} |