Commit
·
94d37f6
1
Parent(s):
ee293ce
Create config.json
Browse files- config.json +34 -0
config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": ["MistralForCausalLM"],
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"hidden_act": "silu",
|
| 6 |
+
"hidden_size": 4096,
|
| 7 |
+
"initializer_range": 0.02,
|
| 8 |
+
"intermediate_size": 14336,
|
| 9 |
+
"max_position_embeddings": 32768,
|
| 10 |
+
"model_type": "mistral",
|
| 11 |
+
"num_attention_heads": 32,
|
| 12 |
+
"num_hidden_layers": 32,
|
| 13 |
+
"num_key_value_heads": 8,
|
| 14 |
+
"rms_norm_eps": 1e-05,
|
| 15 |
+
"rope_theta": 10000.0,
|
| 16 |
+
"sliding_window": 4096,
|
| 17 |
+
"tie_word_embeddings": false,
|
| 18 |
+
"torch_dtype": "bfloat16",
|
| 19 |
+
"transformers_version": "4.34.0.dev0",
|
| 20 |
+
"use_cache": true,
|
| 21 |
+
"vocab_size": 32000,
|
| 22 |
+
|
| 23 |
+
"load_in_4bit": true,
|
| 24 |
+
"bnb_4bit_use_double_quant": true,
|
| 25 |
+
"bnb_4bit_quant_type": "nf4",
|
| 26 |
+
"bnb_4bit_compute_dtype": "torch.bfloat16",
|
| 27 |
+
|
| 28 |
+
"r": 8,
|
| 29 |
+
"lora_alpha": 32,
|
| 30 |
+
"target_modules": ["q_proj", "k_proj"],
|
| 31 |
+
"lora_dropout": 0.05,
|
| 32 |
+
"bias": "none",
|
| 33 |
+
"task_type": "CAUSAL_LM"
|
| 34 |
+
}
|