Upload Phi3ForSequenceClassification
Browse files- config.json +72 -0
- model.safetensors +3 -0
config.json
ADDED
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
|
3 |
+
"architectures": [
|
4 |
+
"Phi3ForSequenceClassification"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"auto_map": {
|
9 |
+
"AutoConfig": "microsoft/Phi-3-mini-4k-instruct--configuration_phi3.Phi3Config",
|
10 |
+
"AutoModelForCausalLM": "microsoft/Phi-3-mini-4k-instruct--modeling_phi3.Phi3ForCausalLM"
|
11 |
+
},
|
12 |
+
"bos_token_id": 1,
|
13 |
+
"embd_pdrop": 0.0,
|
14 |
+
"eos_token_id": 32000,
|
15 |
+
"hidden_act": "silu",
|
16 |
+
"hidden_size": 3072,
|
17 |
+
"id2label": {
|
18 |
+
"0": "age",
|
19 |
+
"1": "disability",
|
20 |
+
"2": "feminine",
|
21 |
+
"3": "general",
|
22 |
+
"4": "masculine",
|
23 |
+
"5": "neutral",
|
24 |
+
"6": "racial",
|
25 |
+
"7": "sexuality"
|
26 |
+
},
|
27 |
+
"initializer_range": 0.02,
|
28 |
+
"intermediate_size": 8192,
|
29 |
+
"label2id": {
|
30 |
+
"age": 0,
|
31 |
+
"disability": 1,
|
32 |
+
"feminine": 2,
|
33 |
+
"general": 3,
|
34 |
+
"masculine": 4,
|
35 |
+
"neutral": 5,
|
36 |
+
"racial": 6,
|
37 |
+
"sexuality": 7
|
38 |
+
},
|
39 |
+
"max_position_embeddings": 4096,
|
40 |
+
"model_type": "phi3",
|
41 |
+
"num_attention_heads": 32,
|
42 |
+
"num_hidden_layers": 32,
|
43 |
+
"num_key_value_heads": 32,
|
44 |
+
"original_max_position_embeddings": 4096,
|
45 |
+
"pad_token_id": 32000,
|
46 |
+
"problem_type": "multi_label_classification",
|
47 |
+
"quantization_config": {
|
48 |
+
"_load_in_4bit": true,
|
49 |
+
"_load_in_8bit": false,
|
50 |
+
"bnb_4bit_compute_dtype": "bfloat16",
|
51 |
+
"bnb_4bit_quant_storage": "uint8",
|
52 |
+
"bnb_4bit_quant_type": "nf4",
|
53 |
+
"bnb_4bit_use_double_quant": true,
|
54 |
+
"llm_int8_enable_fp32_cpu_offload": false,
|
55 |
+
"llm_int8_has_fp16_weight": false,
|
56 |
+
"llm_int8_skip_modules": null,
|
57 |
+
"llm_int8_threshold": 6.0,
|
58 |
+
"load_in_4bit": true,
|
59 |
+
"load_in_8bit": false,
|
60 |
+
"quant_method": "bitsandbytes"
|
61 |
+
},
|
62 |
+
"resid_pdrop": 0.0,
|
63 |
+
"rms_norm_eps": 1e-05,
|
64 |
+
"rope_scaling": null,
|
65 |
+
"rope_theta": 10000.0,
|
66 |
+
"sliding_window": 2047,
|
67 |
+
"tie_word_embeddings": false,
|
68 |
+
"torch_dtype": "float32",
|
69 |
+
"transformers_version": "4.42.4",
|
70 |
+
"use_cache": true,
|
71 |
+
"vocab_size": 32064
|
72 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:092009be748908d1dc912bd35a6c1837b38408935df749a8007ed6f0dfd6a5db
|
3 |
+
size 2264609956
|