Update config.json
Browse files- config.json +37 -13
config.json
CHANGED
@@ -1,16 +1,40 @@
|
|
1 |
{
|
2 |
-
"
|
3 |
-
"
|
4 |
-
"
|
5 |
-
"
|
6 |
-
|
7 |
-
"lora_alpha": 32,
|
8 |
-
"lora_dropout": 0.05,
|
9 |
-
"modules_to_save": null,
|
10 |
-
"peft_type": "LORA",
|
11 |
-
"r": 16,
|
12 |
-
"target_modules": [
|
13 |
-
"query_key_value"
|
14 |
],
|
15 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
}
|
|
|
1 |
{
|
2 |
+
"base_model": "tiiuae/falcon-7b-instruct",
|
3 |
+
"alibi": false,
|
4 |
+
"apply_residual_connection_post_layernorm": false,
|
5 |
+
"architectures": [
|
6 |
+
"RWForCausalLM"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
],
|
8 |
+
"attention_dropout": 0.0,
|
9 |
+
"auto_map": {
|
10 |
+
"AutoConfig": "tiiuae/falcon-7b-instruct--configuration_RW.RWConfig",
|
11 |
+
"AutoModelForCausalLM": "tiiuae/falcon-7b-instruct--modelling_RW.RWForCausalLM"
|
12 |
+
},
|
13 |
+
"bias": false,
|
14 |
+
"bos_token_id": 11,
|
15 |
+
"eos_token_id": 11,
|
16 |
+
"hidden_dropout": 0.0,
|
17 |
+
"hidden_size": 4544,
|
18 |
+
"initializer_range": 0.02,
|
19 |
+
"layer_norm_epsilon": 1e-05,
|
20 |
+
"model_type": "RefinedWebModel",
|
21 |
+
"multi_query": true,
|
22 |
+
"n_head": 71,
|
23 |
+
"n_layer": 32,
|
24 |
+
"parallel_attn": true,
|
25 |
+
"quantization_config": {
|
26 |
+
"bnb_4bit_compute_dtype": "float16",
|
27 |
+
"bnb_4bit_quant_type": "nf4",
|
28 |
+
"bnb_4bit_use_double_quant": false,
|
29 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
30 |
+
"llm_int8_has_fp16_weight": false,
|
31 |
+
"llm_int8_skip_modules": null,
|
32 |
+
"llm_int8_threshold": 6.0,
|
33 |
+
"load_in_4bit": true,
|
34 |
+
"load_in_8bit": false
|
35 |
+
},
|
36 |
+
"torch_dtype": "bfloat16",
|
37 |
+
"transformers_version": "4.30.0.dev0",
|
38 |
+
"use_cache": false,
|
39 |
+
"vocab_size": 65024
|
40 |
}
|