adityamishraog commited on
Commit
29a5649
·
verified ·
1 Parent(s): ee250ec

Upload deepseek-v3-150m-multilingual/checkpoint-84000/config.json with huggingface_hub

Browse files
deepseek-v3-150m-multilingual/checkpoint-84000/config.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DeepseekV3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 2,
8
+ "dtype": "float32",
9
+ "eos_token_id": 3,
10
+ "first_k_dense_replace": 10,
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 600,
14
+ "initializer_range": 0.02,
15
+ "intermediate_size": 1800,
16
+ "kv_lora_rank": 512,
17
+ "max_position_embeddings": 1024,
18
+ "model_type": "deepseek_v3",
19
+ "moe_intermediate_size": 2048,
20
+ "moe_layer_freq": 0,
21
+ "n_group": 8,
22
+ "n_routed_experts": null,
23
+ "n_shared_experts": 1,
24
+ "norm_topk_prob": true,
25
+ "num_attention_heads": 10,
26
+ "num_experts_per_tok": 8,
27
+ "num_hidden_layers": 10,
28
+ "num_key_value_heads": 10,
29
+ "pad_token_id": 0,
30
+ "pretraining_tp": 1,
31
+ "q_lora_rank": 1536,
32
+ "qk_head_dim": 192,
33
+ "qk_nope_head_dim": 128,
34
+ "qk_rope_head_dim": 64,
35
+ "rms_norm_eps": 1e-06,
36
+ "rope_interleave": true,
37
+ "rope_scaling": null,
38
+ "rope_theta": 10000.0,
39
+ "routed_scaling_factor": 2.5,
40
+ "tie_word_embeddings": true,
41
+ "topk_group": 4,
42
+ "transformers_version": "4.56.1",
43
+ "unk_token_id": 1,
44
+ "use_cache": true,
45
+ "v_head_dim": 128,
46
+ "vocab_size": 64000
47
+ }