Hugging Face's logo

mlx-community
/
DeepSeek-V3-4bit

MLX
Safetensors
deepseek_v3
custom_code
4-bit precision
Model card Files Community
1
awni commited on
Commit
ece24eb
·
verified ·
1 Parent(s): 7ffecaa

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. README.md +33 -0
  2. config.json +69 -0
  3. model-00001-of-00088.safetensors +3 -0
  4. model-00002-of-00088.safetensors +3 -0
  5. model-00003-of-00088.safetensors +3 -0
  6. model-00004-of-00088.safetensors +3 -0
  7. model-00005-of-00088.safetensors +3 -0
  8. model-00006-of-00088.safetensors +3 -0
  9. model-00007-of-00088.safetensors +3 -0
  10. model-00008-of-00088.safetensors +3 -0
  11. model-00009-of-00088.safetensors +3 -0
  12. model-00010-of-00088.safetensors +3 -0
  13. model-00011-of-00088.safetensors +3 -0
  14. model-00012-of-00088.safetensors +3 -0
  15. model-00013-of-00088.safetensors +3 -0
  16. model-00014-of-00088.safetensors +3 -0
  17. model-00015-of-00088.safetensors +3 -0
  18. model-00016-of-00088.safetensors +3 -0
  19. model-00017-of-00088.safetensors +3 -0
  20. model-00018-of-00088.safetensors +3 -0
  21. model-00019-of-00088.safetensors +3 -0
  22. model-00020-of-00088.safetensors +3 -0
  23. model-00021-of-00088.safetensors +3 -0
  24. model-00022-of-00088.safetensors +3 -0
  25. model-00023-of-00088.safetensors +3 -0
  26. model-00024-of-00088.safetensors +3 -0
  27. model-00025-of-00088.safetensors +3 -0
  28. model-00026-of-00088.safetensors +3 -0
  29. model-00027-of-00088.safetensors +3 -0
  30. model-00028-of-00088.safetensors +3 -0
  31. model-00029-of-00088.safetensors +3 -0
  32. model-00030-of-00088.safetensors +3 -0
  33. model-00031-of-00088.safetensors +3 -0
  34. model-00032-of-00088.safetensors +3 -0
  35. model-00033-of-00088.safetensors +3 -0
  36. model-00034-of-00088.safetensors +3 -0
  37. model-00035-of-00088.safetensors +3 -0
  38. model-00036-of-00088.safetensors +3 -0
  39. model-00037-of-00088.safetensors +3 -0
  40. model-00038-of-00088.safetensors +3 -0
  41. model-00039-of-00088.safetensors +3 -0
  42. model-00040-of-00088.safetensors +3 -0
  43. model-00041-of-00088.safetensors +3 -0
  44. model-00042-of-00088.safetensors +3 -0
  45. model-00043-of-00088.safetensors +3 -0
  46. model-00044-of-00088.safetensors +3 -0
  47. model-00045-of-00088.safetensors +3 -0
  48. model-00046-of-00088.safetensors +3 -0
  49. model-00047-of-00088.safetensors +3 -0
  50. model-00048-of-00088.safetensors +3 -0
README.md ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: deepseek-ai/DeepSeek-V3
3
+ tags:
4
+ - mlx
5
+ ---
6
+
7
+ # mlx-community/DeepSeek-V3-4bit
8
+
9
+ The Model [mlx-community/DeepSeek-V3-4bit](https://huggingface.co/mlx-community/DeepSeek-V3-4bit) was
10
+ converted to MLX format from [deepseek-ai/DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3)
11
+ using mlx-lm version **0.20.4**.
12
+
13
+ ## Use with mlx
14
+
15
+ ```bash
16
+ pip install mlx-lm
17
+ ```
18
+
19
+ ```python
20
+ from mlx_lm import load, generate
21
+
22
+ model, tokenizer = load("mlx-community/DeepSeek-V3-4bit")
23
+
24
+ prompt = "hello"
25
+
26
+ if tokenizer.chat_template is not None:
27
+ messages = [{"role": "user", "content": prompt}]
28
+ prompt = tokenizer.apply_chat_template(
29
+ messages, add_generation_prompt=True
30
+ )
31
+
32
+ response = generate(model, tokenizer, prompt=prompt, verbose=True)
33
+ ```
config.json ADDED
@@ -0,0 +1,69 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DeepseekV3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_deepseek.DeepseekV3Config",
9
+ "AutoModel": "modeling_deepseek.DeepseekV3Model",
10
+ "AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
11
+ },
12
+ "aux_loss_alpha": 0.001,
13
+ "bos_token_id": 0,
14
+ "eos_token_id": 1,
15
+ "ep_size": 1,
16
+ "first_k_dense_replace": 3,
17
+ "hidden_act": "silu",
18
+ "hidden_size": 7168,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 18432,
21
+ "kv_lora_rank": 512,
22
+ "max_position_embeddings": 163840,
23
+ "model_type": "deepseek_v3",
24
+ "moe_intermediate_size": 2048,
25
+ "moe_layer_freq": 1,
26
+ "n_group": 8,
27
+ "n_routed_experts": 256,
28
+ "n_shared_experts": 1,
29
+ "norm_topk_prob": true,
30
+ "num_attention_heads": 128,
31
+ "num_experts_per_tok": 8,
32
+ "num_hidden_layers": 61,
33
+ "num_key_value_heads": 128,
34
+ "num_nextn_predict_layers": 1,
35
+ "pretraining_tp": 1,
36
+ "q_lora_rank": 1536,
37
+ "qk_nope_head_dim": 128,
38
+ "qk_rope_head_dim": 64,
39
+ "quantization": {
40
+ "group_size": 64,
41
+ "bits": 4
42
+ },
43
+ "quantization_config": {
44
+ "group_size": 64,
45
+ "bits": 4
46
+ },
47
+ "rms_norm_eps": 1e-06,
48
+ "rope_scaling": {
49
+ "beta_fast": 32,
50
+ "beta_slow": 1,
51
+ "factor": 40,
52
+ "mscale": 1.0,
53
+ "mscale_all_dim": 1.0,
54
+ "original_max_position_embeddings": 4096,
55
+ "type": "yarn"
56
+ },
57
+ "rope_theta": 10000,
58
+ "routed_scaling_factor": 2.5,
59
+ "scoring_func": "sigmoid",
60
+ "seq_aux": true,
61
+ "tie_word_embeddings": false,
62
+ "topk_group": 4,
63
+ "topk_method": "noaux_tc",
64
+ "torch_dtype": "bfloat16",
65
+ "transformers_version": "4.33.1",
66
+ "use_cache": true,
67
+ "v_head_dim": 128,
68
+ "vocab_size": 129280
69
+ }
model-00001-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:274e21e8cf16134cf6e25e7e34ba411d79c76b644c06abd7a6495689022c285f
3
+ size 3725148265
model-00002-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd2fc1e876a84dafa897985da8f75bd8358b9e2fc9b556af6f1354f56cec1c83
3
+ size 4361585379
model-00003-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba135b5ad5a9e52bcb0161278a0a6f4b8e016e1f68d846b267d2aaef47aa7fbf
3
+ size 4227859218
model-00004-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49960e6396814c52b33d703e748fa3ae690376236d29961b74840c8dbf74627e
3
+ size 4361585451
model-00005-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d7883c8433e9c93600bf4858af545b01c7c90af4fdfd8f986c399b57420dda6
3
+ size 4361585389
model-00006-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:75133be469b5f560eadcb3e203c4b43ceedf28a6cee3433f1c6fcacadf925949
3
+ size 4227859214
model-00007-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d66ef82bd89fe180b86394980b24fd625a79159a4a9697babbc18ad1185c0a18
3
+ size 4361585449
model-00008-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a6e6388c10a932a20ff45e4a973de95ea0f19591fbd89d58695b873e106cd52
3
+ size 4361585357
model-00009-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f0d1760bc8aaea9c7188f0d6c2e3b29fe45ada27382eb4ba0cc721294ca174a
3
+ size 4227859218
model-00010-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a48826202563cba72cc2dfdc5433b0ab3c3e34d4b9b5ccbba53fda2028619970
3
+ size 4361585375
model-00011-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2ecb965051e6894865b5354cc153d5252b3a4d8a02dec734cad7b9afeb26afb
3
+ size 4361585378
model-00012-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fcc6de77e8d2e42ea8fddbf67652d73f900c172570d26ac2abe0aa2daf4698ae
3
+ size 4227859224
model-00013-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dde42e965cbb807d3c658347540744ddde4cd800d907dc2d75b485ac0bc240d4
3
+ size 4361585461
model-00014-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4fdd489ee9596ad3f1a16b3768cad22365782cbfeac2d5486965b56a2b20847
3
+ size 4361585391
model-00015-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:561b25ab9f01ef8a21c8a77afa57edddca61f3f6f405fdb08a33315bd7a83003
3
+ size 4227859224
model-00016-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b9714b18a771df44f482f7e1f9489ece51aae65dea8eaff56025188f57ebcb3c
3
+ size 4361585455
model-00017-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a132b8c642002f4ecca1d2e4bc6cd2bf6609b0312351418c344ff7c04d8a32f6
3
+ size 4361585389
model-00018-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd2b5f731322193e8f2a871f9a99731cebdce7047e65e8988d14363560bf0e1c
3
+ size 4227859220
model-00019-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a69996734871b7795c6a07108c30233f1441a95b13305650d79b0741620aeaad
3
+ size 4361585463
model-00020-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f812e89ee805d02accbba31e179d116c275fda624f9aaac647ebf6f92426d6bb
3
+ size 4361585423
model-00021-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3bded52b9128955809b03f70ce6f4c8bb6f54552c93e8fa72f9fcd3e9d80f96d
3
+ size 4227859224
model-00022-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49fc4be1da8c0913ed64294685d27a754e02657c68d9b8f13650b391408a3b81
3
+ size 4361585441
model-00023-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6573098f95beb56513c4dac7b3b16635a8f054f0b41028d329956efaee56111
3
+ size 4361585399
model-00024-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6a7c80bade073eb077729eb4f335a8fe961f6a20145f061b9b68ac8a1c3a0349
3
+ size 4227859226
model-00025-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e3b93707952d2411e21c2be6bfa9742bca771a79e34fa90f24c9243321cc67f
3
+ size 4361585485
model-00026-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4e3fd43968a4d91cc4d31c55646225177722f88b8419b4b549c483690aa123c
3
+ size 4361585403
model-00027-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a58d235dd48a27de00a2ab7cf50ad275f45d6331321472cfbd55c0f10324691
3
+ size 4227859220
model-00028-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:73e52aafbacc21506d61184177797d4b7dd2da6033fa3059872cf31d07823cb6
3
+ size 4361585453
model-00029-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:17356814f3bfbba6748226f22e5f3845703b0648e1d3d30f9e717e983a4fb46d
3
+ size 4361585371
model-00030-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:99ec3065e8f776236c8bb5e6bbf3c7dad16903efdb9e4633ef7eedc76a49a667
3
+ size 4227859222
model-00031-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f6d2039068f13738db96142f731907cbdedeadb09735d4c8bfbf3484b2003e6
3
+ size 4361585487
model-00032-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3eb57c29fe3ef6e6c23a36c2660ffac48eb8f908e03ddd86fe2616913a96fb29
3
+ size 4361585433
model-00033-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea58586fc8f021d662b6332427d7c764586ed9c137977d338aa785c9a8b105a1
3
+ size 4227859218
model-00034-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aed685ac7d36cb790d2d94160e6edecc743ac71d90e5d7e8f76c481977f30231
3
+ size 4361585441
model-00035-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11387a9f103e362366843c44c40e7f04914c4c337df3fed3142cc7bbf7526ef1
3
+ size 4361585411
model-00036-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37d324239cd367fff09e1d51cf7eef95507616df3771074d9377bddf349b4ae6
3
+ size 4227859224
model-00037-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06bae7efc92f661aa4e9c3dbaf37198414d70effd31277c92270c94f77a78346
3
+ size 4361585487
model-00038-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:399d275ee83c7f686f3580f182be9cc27e7493c0a52d8c25596dbc9ba062c454
3
+ size 4361585399
model-00039-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:220ffbfe905d6190edd0db889d34e5828b16e6f649ca3a85b3bde8990adeacc1
3
+ size 4227859220
model-00040-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6c8852f7c16a3678ff855bbad972bdea9bd31a7f81816fd3f9973e103bd0c5b
3
+ size 4361585427
model-00041-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a980c421bdcf805a8ddd6d0be21a9851f39421bccb1d263f5d6abcdaa540c2fa
3
+ size 4361585389
model-00042-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1538d9bcadcb54d97ea248364c3275d22b8f846f058332c84ac49bb911ba584
3
+ size 4227859224
model-00043-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db3921c5ca0878b4faf4fb406f47b3f82db6812f99183e0bdf6666c86d6e0fad
3
+ size 4361585485
model-00044-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b76cb515ae11b6fe7217e81a9a69e1f337ac8110f6fbfcc5160e095b205754c
3
+ size 4361585425
model-00045-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:52342f26956a1cee607ea754b47cd0a74bd0454ce394cb33dce40f5e69fd8ac1
3
+ size 4227859224
model-00046-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4291c42ad658a6e139470d1806426090ba6c1d9a351d3640be5c25bec67efd5
3
+ size 4361585485
model-00047-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d42ad7636fe3a2b4d25489ff5c67fcf2ac746a254e1e62e120ecbcedf35f197
3
+ size 4361585385
model-00048-of-00088.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f600711bff9b8289f729c95bebf3217069eb55890f306f84ee784d9c3a5d7d9
3
+ size 4227859224