31 lines
875 B
JSON
31 lines
875 B
JSON
|
|
{
|
||
|
|
"_name_or_path": "/scratch/tikhomirov/workdir/data/models/2025/qwen/14b/ruadapt_qwen2.5_14B_part1-2_lr1e4_wsd_bs128_v2_peft_32_32_lm_head_1e4_v7_cot_r1_v1/ft0",
|
||
|
|
"architectures": [
|
||
|
|
"Qwen2ForCausalLM"
|
||
|
|
],
|
||
|
|
"attention_dropout": 0.0,
|
||
|
|
"bos_token_id": 145157,
|
||
|
|
"eos_token_id": 145155,
|
||
|
|
"hidden_act": "silu",
|
||
|
|
"hidden_size": 5120,
|
||
|
|
"initializer_range": 0.02,
|
||
|
|
"intermediate_size": 13824,
|
||
|
|
"max_position_embeddings": 131072,
|
||
|
|
"max_window_layers": 48,
|
||
|
|
"model_type": "qwen2",
|
||
|
|
"num_attention_heads": 40,
|
||
|
|
"num_hidden_layers": 48,
|
||
|
|
"num_key_value_heads": 8,
|
||
|
|
"pad_token_id": 145155,
|
||
|
|
"rms_norm_eps": 1e-05,
|
||
|
|
"rope_scaling": null,
|
||
|
|
"rope_theta": 1000000.0,
|
||
|
|
"sliding_window": null,
|
||
|
|
"tie_word_embeddings": false,
|
||
|
|
"torch_dtype": "bfloat16",
|
||
|
|
"transformers_version": "4.45.2",
|
||
|
|
"use_cache": true,
|
||
|
|
"use_sliding_window": false,
|
||
|
|
"vocab_size": 145159
|
||
|
|
}
|