ltoniazzi commited on
Commit
e956224
·
verified ·
1 Parent(s): 8c62b16

Upload Phi3ForCausalLM/hidden_size=64/base/config.json with huggingface_hub

Browse files
Phi3ForCausalLM/hidden_size=64/base/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "models/Phi3ForCausalLM_hidden_size=64/base_untrained",
3
+ "architectures": [
4
+ "Phi3ForCausalLM"
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": 64,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 64,
19
+ "max_position_embeddings": 4096,
20
+ "model_type": "phi3",
21
+ "num_attention_heads": 8,
22
+ "num_hidden_layers": 32,
23
+ "num_key_value_heads": 8,
24
+ "original_max_position_embeddings": 4096,
25
+ "pad_token_id": 32000,
26
+ "resid_pdrop": 0.0,
27
+ "rms_norm_eps": 1e-05,
28
+ "rope_scaling": null,
29
+ "rope_theta": 10000.0,
30
+ "sliding_window": 2047,
31
+ "tie_word_embeddings": false,
32
+ "torch_dtype": "float32",
33
+ "transformers_version": "4.44.0",
34
+ "use_cache": false,
35
+ "vocab_size": 32064
36
+ }