gev-e2b / gev.json
onatm's picture
Upload folder using huggingface_hub
b15926c verified
Raw History Blame Contribute Delete
2.07 kB
{
"format": "gev-checkpoint",
"format_version": 1,
"gev_version": "0.1.0",
"base_model": "google/gemma-4-E2B",
"revision": "d29ff6b45f081a49ee2733a859c9c9c2d95d1a6f",
"family": "gemma4",
"markers": {
"state": {
"token": "<unused0>",
"id": 6
},
"question": {
"token": "<unused1>",
"id": 7
},
"option_start": {
"token": "<unused2>",
"id": 8
},
"option_end": {
"token": "<unused3>",
"id": 9
},
"decide": {
"token": "<unused4>",
"id": 10
}
},
"state_cap": 384,
"branch_cap": 2048,
"temperature": 1.6245047927124707,
"trained_with": {
"backend": "mlx",
"dtype": "bf16"
},
"config": {
"name": "gev-e2b",
"model": {
"name": "google/gemma-4-E2B",
"revision": "d29ff6b45f081a49ee2733a859c9c9c2d95d1a6f",
"family": "gemma4",
"markers": [
"<unused0>",
"<unused1>",
"<unused2>",
"<unused3>",
"<unused4>"
],
"lora_rank": 16,
"lora_alpha": 32,
"lora_dropout": 0.05,
"lora_targets": [
"q_proj",
"k_proj",
"v_proj",
"o_proj",
"gate_proj",
"up_proj",
"down_proj"
],
"pointer_width": 256
},
"training": {
"seed": 0,
"epochs": 2,
"learning_rate": 0.0001,
"head_learning_rate": null,
"weight_decay": 0.01,
"logical_batch": 8,
"microbatch": 1,
"max_steps": null,
"save_every": 250,
"state_cap": 384,
"branch_cap": 2048,
"p_none": 0.1,
"p_none_distract": 0.12,
"p_distract": 0.15,
"p_none_pair": 0.25
},
"backend": "mlx",
"device": "auto",
"dtype": "bf16",
"attn_implementation": "sdpa",
"gradient_checkpointing": false
},
"training": {
"steps": 3144,
"total_steps": 3144,
"complete": true,
"seconds": 13010.148587625008,
"train_sha256": "7ed5254b5cb5291baefaceb09edf7e13110258211518c8038f4a12c11bd628ad",
"train_records": 12576
}
}