uploaded model
Browse files- config.json +61 -0
config.json
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "mpt_1000_STEPS_1e6_SFT_SFT",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MptForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attn_config": {
|
| 7 |
+
"model_type": ""
|
| 8 |
+
},
|
| 9 |
+
"auto_map": {
|
| 10 |
+
"AutoConfig": "mosaicml/mpt-7b-instruct--configuration_mpt.MPTConfig",
|
| 11 |
+
"AutoModelForCausalLM": "mosaicml/mpt-7b-instruct--modeling_mpt.MPTForCausalLM"
|
| 12 |
+
},
|
| 13 |
+
"d_model": 4096,
|
| 14 |
+
"emb_pdrop": 0,
|
| 15 |
+
"embedding_fraction": 1.0,
|
| 16 |
+
"expansion_ratio": 4,
|
| 17 |
+
"init_config": {
|
| 18 |
+
"emb_init_std": null,
|
| 19 |
+
"emb_init_uniform_lim": null,
|
| 20 |
+
"fan_mode": "fan_in",
|
| 21 |
+
"init_div_is_residual": true,
|
| 22 |
+
"init_gain": 0,
|
| 23 |
+
"init_nonlinearity": "relu",
|
| 24 |
+
"init_std": 0.02,
|
| 25 |
+
"name": "kaiming_normal_",
|
| 26 |
+
"verbose": 0
|
| 27 |
+
},
|
| 28 |
+
"init_device": "cpu",
|
| 29 |
+
"initializer_range": 0.02,
|
| 30 |
+
"layer_norm_epsilon": 1e-05,
|
| 31 |
+
"learned_pos_emb": true,
|
| 32 |
+
"logit_scale": null,
|
| 33 |
+
"max_seq_len": 2048,
|
| 34 |
+
"model_type": "mpt",
|
| 35 |
+
"n_heads": 32,
|
| 36 |
+
"n_layers": 32,
|
| 37 |
+
"no_bias": true,
|
| 38 |
+
"norm_type": "low_precision_layernorm",
|
| 39 |
+
"quantization_config": {
|
| 40 |
+
"_load_in_4bit": false,
|
| 41 |
+
"_load_in_8bit": true,
|
| 42 |
+
"bnb_4bit_compute_dtype": "float32",
|
| 43 |
+
"bnb_4bit_quant_storage": "uint8",
|
| 44 |
+
"bnb_4bit_quant_type": "fp4",
|
| 45 |
+
"bnb_4bit_use_double_quant": false,
|
| 46 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
| 47 |
+
"llm_int8_has_fp16_weight": false,
|
| 48 |
+
"llm_int8_skip_modules": null,
|
| 49 |
+
"llm_int8_threshold": 6.0,
|
| 50 |
+
"load_in_4bit": false,
|
| 51 |
+
"load_in_8bit": true,
|
| 52 |
+
"quant_method": "bitsandbytes"
|
| 53 |
+
},
|
| 54 |
+
"resid_pdrop": 0,
|
| 55 |
+
"tokenizer_name": "EleutherAI/gpt-neox-20b",
|
| 56 |
+
"torch_dtype": "float16",
|
| 57 |
+
"transformers_version": "4.47.1",
|
| 58 |
+
"use_cache": false,
|
| 59 |
+
"verbose": 0,
|
| 60 |
+
"vocab_size": 50432
|
| 61 |
+
}
|