T-tchi0 commited on
Commit
6ba7fd4
·
verified ·
1 Parent(s): 8eb706c

Upload PhiForCausalLM

Browse files
Files changed (3) hide show
  1. config.json +7 -7
  2. generation_config.json +0 -2
  3. model.safetensors +2 -2
config.json CHANGED
@@ -3,20 +3,20 @@
3
  "PhiForCausalLM"
4
  ],
5
  "attention_dropout": 0.0,
6
- "bos_token_id": 50256,
7
  "embd_pdrop": 0.0,
8
- "eos_token_id": 50256,
9
  "hidden_act": "gelu_new",
10
- "hidden_size": 2560,
11
  "initializer_range": 0.02,
12
- "intermediate_size": 10240,
13
  "layer_norm_eps": 1e-05,
14
  "max_position_embeddings": 2048,
15
  "model_type": "phi",
16
  "num_attention_heads": 32,
17
- "num_hidden_layers": 32,
18
  "num_key_value_heads": 32,
19
- "partial_rotary_factor": 0.4,
20
  "qk_layernorm": false,
21
  "quantization_config": {
22
  "_load_in_4bit": true,
@@ -33,7 +33,7 @@
33
  "load_in_8bit": false,
34
  "quant_method": "bitsandbytes"
35
  },
36
- "resid_pdrop": 0.1,
37
  "rope_scaling": null,
38
  "rope_theta": 10000.0,
39
  "tie_word_embeddings": false,
 
3
  "PhiForCausalLM"
4
  ],
5
  "attention_dropout": 0.0,
6
+ "bos_token_id": null,
7
  "embd_pdrop": 0.0,
8
+ "eos_token_id": null,
9
  "hidden_act": "gelu_new",
10
+ "hidden_size": 2048,
11
  "initializer_range": 0.02,
12
+ "intermediate_size": 8192,
13
  "layer_norm_eps": 1e-05,
14
  "max_position_embeddings": 2048,
15
  "model_type": "phi",
16
  "num_attention_heads": 32,
17
+ "num_hidden_layers": 24,
18
  "num_key_value_heads": 32,
19
+ "partial_rotary_factor": 0.5,
20
  "qk_layernorm": false,
21
  "quantization_config": {
22
  "_load_in_4bit": true,
 
33
  "load_in_8bit": false,
34
  "quant_method": "bitsandbytes"
35
  },
36
+ "resid_pdrop": 0.0,
37
  "rope_scaling": null,
38
  "rope_theta": 10000.0,
39
  "tie_word_embeddings": false,
generation_config.json CHANGED
@@ -1,6 +1,4 @@
1
  {
2
  "_from_model_config": true,
3
- "bos_token_id": 50256,
4
- "eos_token_id": 50256,
5
  "transformers_version": "4.51.3"
6
  }
 
1
  {
2
  "_from_model_config": true,
 
 
3
  "transformers_version": "4.51.3"
4
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ec0332d7d5af2984770ce588cfd00645de6b007fca329ca48547afb47951f337
3
- size 1824834112
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:344959645c80123c85535e4ae0db224057aed57b99b359afd031a1b5e7907a54
3
+ size 1044073644