laya-banking77 / rl_agent_config.json
VibeHash's picture
Add laya-banking77: Laya fine-tuned on Banking77, its scores and the code that made it
4e09985 verified
Raw History Blame Contribute Delete
5.06 kB
{
"encoder": "answerdotai/ModernBERT-large",
"head_layers": 2,
"max_len": 512,
"head_max_len": 512,
"max_prefixes": 6,
"act_costs": {
"escalate": 0.5
},
"cost_wrong_act": 3.0,
"amp_dtype": "bf16",
"model_name": "rl-agent",
"temperature": [
3.49420428276062,
1.0,
1.0
],
"training": {
"updates": 7313,
"epochs_completed": 1,
"hours": 1.96,
"world_size": 1,
"fine_tuned_from_checkpoint": true,
"laya_train": {
"epochs": 4,
"micro_batch": 8,
"grad_accum": 8,
"encoder_lr": 2.5e-05,
"head_lr": 0.0001,
"min_lr": 1e-06,
"weight_decay": 0.01,
"grad_clip": 1.0,
"loss": "rlcd",
"rl_samples": 4,
"sigma_start": 0.4,
"sigma_end": 0.1,
"w_sph": 0.75,
"w_rps": 1.0,
"shuffle_options": [],
"option_layout": null,
"freeze_encoder": false,
"calib_max": 400,
"calib_frac": 0.1,
"calib_seed": 20260922,
"seed": 0,
"max_len": null,
"head_max_len": 512,
"amp": null,
"gradient_checkpointing": null,
"log_every": 100,
"label_smoothing": 0.0,
"text_column": "text",
"label_column": "label",
"question_id": "label",
"instructions": null,
"eval_data": null,
"target_error": 0.01,
"min_abstain_n": 10
},
"laya_train_calibration": {
"choice": {
"items": 400,
"temperature": 3.49420428276062,
"issues": [],
"accuracy": 0.9075,
"mean_confidence": 0.9215655326454519
},
"score": {
"items": 0,
"temperature": 1.0,
"issues": []
},
"noul": {
"items": 0,
"temperature": 1.0,
"issues": []
}
},
"train_report": {
"eval_source": "calibration_slice",
"eval_mode": "in_sample_calibration",
"is_held_out": false,
"eval_items": 400,
"note": "Metrics reflect in-sample calibration fit, not independent held-out evaluation. Provide --eval for held-out validation.",
"before": {
"items": 400,
"loss": 11.0271,
"accuracy": 0.4375,
"mean_confidence": 0.8994,
"ece": 0.4665,
"brier": 1.0062,
"brier_top1": 0.4425,
"by_type": {
"choice": {
"items": 400,
"loss": 11.0271,
"accuracy": 0.4375,
"mean_confidence": 0.8994,
"ece": 0.4665,
"brier": 1.0062,
"brier_top1": 0.4425
}
}
},
"after": {
"items": 400,
"loss": 0.4723,
"accuracy": 0.9075,
"mean_confidence": 0.9216,
"ece": 0.0399,
"brier": 0.1525,
"brier_top1": 0.0651,
"by_type": {
"choice": {
"items": 400,
"loss": 0.4723,
"accuracy": 0.9075,
"mean_confidence": 0.9216,
"ece": 0.0399,
"brier": 0.1525,
"brier_top1": 0.0651
}
}
},
"comparison": {
"delta_accuracy": 0.47,
"delta_loss": -10.5548,
"delta_ece": -0.4266,
"delta_brier": -0.8537,
"delta_brier_top1": -0.3774,
"delta_mean_confidence": 0.0222
},
"calibration": {
"choice": {
"items": 400,
"temperature": 3.49420428276062,
"issues": [],
"accuracy": 0.9075,
"mean_confidence": 0.9215655326454519
},
"score": {
"items": 0,
"temperature": 1.0,
"issues": []
},
"noul": {
"items": 0,
"temperature": 1.0,
"issues": []
}
},
"abstention_thresholds": {
"choice:11+": 0.979
},
"training": {
"epochs": 4,
"micro_batch": 8,
"grad_accum": 8,
"encoder_lr": 2.5e-05,
"head_lr": 0.0001,
"min_lr": 1e-06,
"weight_decay": 0.01,
"grad_clip": 1.0,
"loss": "rlcd",
"rl_samples": 4,
"sigma_start": 0.4,
"sigma_end": 0.1,
"w_sph": 0.75,
"w_rps": 1.0,
"shuffle_options": [],
"option_layout": null,
"freeze_encoder": false,
"calib_max": 400,
"calib_frac": 0.1,
"calib_seed": 20260922,
"seed": 0,
"max_len": null,
"head_max_len": 512,
"amp": null,
"gradient_checkpointing": null,
"log_every": 100,
"label_smoothing": 0.0,
"text_column": "text",
"label_column": "label",
"question_id": "label",
"instructions": null,
"eval_data": null,
"target_error": 0.01,
"min_abstain_n": 10
}
},
"abstention_thresholds": {
"choice:11+": 0.979
}
},
"fine_tuned": true
}