danielhanchen commited on
Commit
af6c5d4
1 Parent(s): eecad33

Upload Qwen2ForCausalLM

Browse files
config.json CHANGED
@@ -5,7 +5,7 @@
5
  ],
6
  "attention_dropout": 0.0,
7
  "bos_token_id": 151643,
8
- "eos_token_id": 151643,
9
  "hidden_act": "silu",
10
  "hidden_size": 3584,
11
  "initializer_range": 0.02,
@@ -16,6 +16,7 @@
16
  "num_attention_heads": 28,
17
  "num_hidden_layers": 28,
18
  "num_key_value_heads": 4,
 
19
  "quantization_config": {
20
  "_load_in_4bit": true,
21
  "_load_in_8bit": false,
@@ -37,6 +38,7 @@
37
  "tie_word_embeddings": false,
38
  "torch_dtype": "bfloat16",
39
  "transformers_version": "4.44.2",
 
40
  "use_cache": true,
41
  "use_sliding_window": false,
42
  "vocab_size": 152064
 
5
  ],
6
  "attention_dropout": 0.0,
7
  "bos_token_id": 151643,
8
+ "eos_token_id": 151645,
9
  "hidden_act": "silu",
10
  "hidden_size": 3584,
11
  "initializer_range": 0.02,
 
16
  "num_attention_heads": 28,
17
  "num_hidden_layers": 28,
18
  "num_key_value_heads": 4,
19
+ "pad_token_id": 151665,
20
  "quantization_config": {
21
  "_load_in_4bit": true,
22
  "_load_in_8bit": false,
 
38
  "tie_word_embeddings": false,
39
  "torch_dtype": "bfloat16",
40
  "transformers_version": "4.44.2",
41
+ "unsloth_fixed": true,
42
  "use_cache": true,
43
  "use_sliding_window": false,
44
  "vocab_size": 152064
generation_config.json CHANGED
@@ -5,7 +5,8 @@
5
  151645,
6
  151643
7
  ],
8
- "pad_token_id": 151643,
 
9
  "repetition_penalty": 1.1,
10
  "temperature": 0.7,
11
  "top_k": 20,
 
5
  151645,
6
  151643
7
  ],
8
+ "max_length": 32768,
9
+ "pad_token_id": 151665,
10
  "repetition_penalty": 1.1,
11
  "temperature": 0.7,
12
  "top_k": 20,
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f752ce6268bad52b1c38c41569b1b6b8fd5b43edc51abcbca40f13c1db8690e6
3
+ size 4457259784
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5aa6e5cbe642377fd441fb4e60e83cca96b2bcd9820e245b9ea06d94653f17f2
3
+ size 1089994880
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff