Hugging Face's logo

JingzeShi
/
OpenSeek-1.4B-A0.4B

Text Generation
Transformers
Safetensors
deepseek_v3
conversational
text-generation-inference
Model card Files
xet
Community
JingzeShi commited on
Commit
ab5430f
·
verified ·
1 Parent(s): b9fbfab

Upload DeepseekV3ForCausalLM

Browse files
Files changed (3) hide show
  1. config.json +51 -47
  2. generation_config.json +6 -6
  3. model.safetensors +2 -2
config.json CHANGED
@@ -1,47 +1,51 @@
1
- {
2
- "architectures": [
3
- "DeepseekV3ForCausalLM"
4
- ],
5
- "attention_bias": false,
6
- "attention_dropout": 0.0,
7
- "bos_token_id": 0,
8
- "eos_token_id": 1,
9
- "first_k_dense_replace": 1,
10
- "head_dim": 64,
11
- "hidden_act": "silu",
12
- "hidden_size": 1280,
13
- "initializer_range": 0.006,
14
- "intermediate_size": 7168,
15
- "kv_lora_rank": 512,
16
- "max_position_embeddings": 4096,
17
- "model_type": "openseek",
18
- "moe_intermediate_size": 2048,
19
- "n_group": 8,
20
- "n_routed_experts": 256,
21
- "n_shared_experts": 1,
22
- "norm_topk_prob": false,
23
- "num_attention_heads": 10,
24
- "num_experts": 78400,
25
- "num_experts_per_tok": 16,
26
- "num_hidden_layers": 6,
27
- "num_key_value_heads": 10,
28
- "output_router_logits": false,
29
- "pretraining_tp": 1,
30
- "q_lora_rank": null,
31
- "qk_head_dim": 192,
32
- "qk_nope_head_dim": 128,
33
- "qk_rope_head_dim": 64,
34
- "rms_norm_eps": 1e-06,
35
- "rope_interleave": true,
36
- "rope_scaling": null,
37
- "rope_theta": 1000000,
38
- "routed_scaling_factor": 2.5,
39
- "router_aux_loss_coef": 0.001,
40
- "tie_word_embeddings": true,
41
- "topk_group": 4,
42
- "torch_dtype": "bfloat16",
43
- "transformers_version": "4.55.0",
44
- "use_cache": true,
45
- "v_head_dim": 128,
46
- "vocab_size": 151851
47
- }
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DeepseekV3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "aux_loss_alpha": 0.001,
8
+ "bos_token_id": 0,
9
+ "eos_token_id": 1,
10
+ "ep_size": 1,
11
+ "first_k_dense_replace": 1,
12
+ "head_dim": 64,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 1280,
15
+ "initializer_range": 0.006,
16
+ "intermediate_size": 7168,
17
+ "kv_lora_rank": 512,
18
+ "max_position_embeddings": 4096,
19
+ "model_type": "deepseek_v3",
20
+ "moe_intermediate_size": 896,
21
+ "moe_layer_freq": 1,
22
+ "n_group": 1,
23
+ "n_routed_experts": 64,
24
+ "n_shared_experts": 2,
25
+ "norm_topk_prob": true,
26
+ "num_attention_heads": 10,
27
+ "num_experts_per_tok": 6,
28
+ "num_hidden_layers": 6,
29
+ "num_key_value_heads": 10,
30
+ "num_nextn_predict_layers": 1,
31
+ "pretraining_tp": 1,
32
+ "q_lora_rank": null,
33
+ "qk_head_dim": 192,
34
+ "qk_nope_head_dim": 128,
35
+ "qk_rope_head_dim": 64,
36
+ "rms_norm_eps": 1e-06,
37
+ "rope_interleave": true,
38
+ "rope_scaling": null,
39
+ "rope_theta": 1000000,
40
+ "routed_scaling_factor": 2.446,
41
+ "scoring_func": "sigmoid",
42
+ "seq_aux": true,
43
+ "tie_word_embeddings": true,
44
+ "topk_group": 1,
45
+ "topk_method": "noaux_tc",
46
+ "torch_dtype": "bfloat16",
47
+ "transformers_version": "4.55.0",
48
+ "use_cache": true,
49
+ "v_head_dim": 128,
50
+ "vocab_size": 151851
51
+ }
generation_config.json CHANGED
@@ -1,6 +1,6 @@
1
- {
2
- "_from_model_config": true,
3
- "bos_token_id": 0,
4
- "eos_token_id": 1,
5
- "transformers_version": "4.55.0"
6
- }
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 0,
4
+ "eos_token_id": 1,
5
+ "transformers_version": "4.55.0"
6
+ }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1c1e999761dbc8c13e48eef3bd4ec73005bd8a8a54dc094861c99b737740e192
3
- size 2819824912
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20ca47d8af781862049e38ca46c12dd04590befc3f991e01e79e6793928080c8
3
+ size 2789324832