Files changed (3) hide show
  1. config.json +7 -4
  2. generation_config.json +1 -1
  3. model.safetensors +2 -2
config.json CHANGED
@@ -7,7 +7,7 @@
7
  "attention_dropout": 0.0,
8
  "bos_token_id": 2,
9
  "cache_implementation": "hybrid",
10
- "dtype": "bfloat16",
11
  "eos_token_id": 3,
12
  "head_dim": 128,
13
  "hidden_act": "silu",
@@ -29,12 +29,15 @@
29
  "order_of_interleaved_layers": "local_attn_first",
30
  "pad_token_id": 0,
31
  "position_embedding_type": "rope_gptj",
32
- "rope_scaling": null,
33
- "rope_theta": 50000,
 
 
34
  "rotary_pct": 1.0,
35
  "sliding_window": 4096,
36
  "sliding_window_pattern": 4,
37
- "transformers_version": "4.57.0",
 
38
  "use_cache": true,
39
  "use_embedding_sharing": true,
40
  "use_gated_activation": true,
 
7
  "attention_dropout": 0.0,
8
  "bos_token_id": 2,
9
  "cache_implementation": "hybrid",
10
+ "dtype": "float32",
11
  "eos_token_id": 3,
12
  "head_dim": 128,
13
  "hidden_act": "silu",
 
29
  "order_of_interleaved_layers": "local_attn_first",
30
  "pad_token_id": 0,
31
  "position_embedding_type": "rope_gptj",
32
+ "rope_parameters": {
33
+ "rope_theta": 50000,
34
+ "rope_type": "default"
35
+ },
36
  "rotary_pct": 1.0,
37
  "sliding_window": 4096,
38
  "sliding_window_pattern": 4,
39
+ "tie_word_embeddings": true,
40
+ "transformers_version": "5.0.0",
41
  "use_cache": true,
42
  "use_embedding_sharing": true,
43
  "use_gated_activation": true,
generation_config.json CHANGED
@@ -4,5 +4,5 @@
4
  "cache_implementation": "hybrid",
5
  "eos_token_id": 3,
6
  "pad_token_id": 0,
7
- "transformers_version": "4.57.0"
8
  }
 
4
  "cache_implementation": "hybrid",
5
  "eos_token_id": 3,
6
  "pad_token_id": 0,
7
+ "transformers_version": "5.0.0"
8
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2adbd74f6bef3d7ee02a8ebecd2547b0a83e2115359226643b280c19ea45424f
3
- size 2132072
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae21f5d6adb86bc9727cfce416f984073645216d0a9e390a3637bc17fbe93e51
3
+ size 4262232