JoseCarlosGarcia95 commited on
Commit
825f40a
·
1 Parent(s): f6f7c43

replace config files

Browse files

Signed-off-by: José Carlos García <[email protected]>

Files changed (2) hide show
  1. config.json +9 -7
  2. generation_config.json +1 -1
config.json CHANGED
@@ -1,23 +1,25 @@
1
  {
2
- "_name_or_path": "/home/ec2-user/checkpoints/falcon3-1b-1bit/hf-1bit",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
6
  "attention_bias": false,
7
  "attention_dropout": 0.0,
 
8
  "eos_token_id": 11,
9
  "head_dim": 256,
10
  "hidden_act": "silu",
11
- "hidden_size": 2048,
12
  "initializer_range": 0.02,
13
- "intermediate_size": 8192,
14
  "is_bitnet_config": true,
15
- "max_position_embeddings": 8192,
16
  "mlp_bias": false,
17
  "model_type": "llama",
18
- "num_attention_heads": 8,
19
- "num_hidden_layers": 18,
20
  "num_key_value_heads": 4,
 
21
  "pretraining_tp": 1,
22
  "quantization_config": {
23
  "modules_to_not_convert": [
@@ -30,7 +32,7 @@
30
  "rope_theta": 1000042,
31
  "tie_word_embeddings": false,
32
  "torch_dtype": "bfloat16",
33
- "transformers_version": "4.48.0.dev0",
34
  "use_cache": true,
35
  "vocab_size": 131072
36
  }
 
1
  {
2
+ "activation": "swiglu",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
6
  "attention_bias": false,
7
  "attention_dropout": 0.0,
8
+ "bos_token_id": 11,
9
  "eos_token_id": 11,
10
  "head_dim": 256,
11
  "hidden_act": "silu",
12
+ "hidden_size": 3072,
13
  "initializer_range": 0.02,
14
+ "intermediate_size": 9216,
15
  "is_bitnet_config": true,
16
+ "max_position_embeddings": 4096,
17
  "mlp_bias": false,
18
  "model_type": "llama",
19
+ "num_attention_heads": 12,
20
+ "num_hidden_layers": 22,
21
  "num_key_value_heads": 4,
22
+ "parallel_attn": false,
23
  "pretraining_tp": 1,
24
  "quantization_config": {
25
  "modules_to_not_convert": [
 
32
  "rope_theta": 1000042,
33
  "tie_word_embeddings": false,
34
  "torch_dtype": "bfloat16",
35
+ "transformers_version": "4.46.0.dev0",
36
  "use_cache": true,
37
  "vocab_size": 131072
38
  }
generation_config.json CHANGED
@@ -2,5 +2,5 @@
2
  "_from_model_config": true,
3
  "bos_token_id": 11,
4
  "eos_token_id": 11,
5
- "transformers_version": "4.48.0.dev0"
6
  }
 
2
  "_from_model_config": true,
3
  "bos_token_id": 11,
4
  "eos_token_id": 11,
5
+ "transformers_version": "4.46.0.dev0"
6
  }