Hugging Face's logo

Sigfinity
/
crystal-math-gpt-oss-20b

Text Generation
Transformers
Safetensors
English
gpt_oss
text-generation-inference
unsloth
conversational
4-bit precision
bitsandbytes
Model card Files
xet
 Community
whalexdfsa commited on
Commit
95993d5
·
verified ·
1 Parent(s): 6e483cb

(Trained with Unsloth)

Browse files
Files changed (6) hide show
  1. config.json +83 -68
  2. generation_config.json +11 -0
  3. model-00001-of-00003.safetensors +3 -0
  4. model-00002-of-00003.safetensors +3 -0
  5. model-00003-of-00003.safetensors +3 -0
  6. model.safetensors.index.json +0 -0
config.json CHANGED
@@ -1,70 +1,85 @@
1
  {
2
- "architectures": [
3
- "GptOssForCausalLM"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  ],
5
- "attention_bias": true,
6
- "attention_dropout": 0.0,
7
- "bos_token_id": 199998,
8
- "torch_dtype": "bfloat16",
9
- "eos_token_id": 200002,
10
- "experts_per_token": 4,
11
- "head_dim": 64,
12
- "hidden_act": "silu",
13
- "hidden_size": 2880,
14
- "initial_context_length": 4096,
15
- "initializer_range": 0.02,
16
- "intermediate_size": 2880,
17
- "layer_types": [
18
- "sliding_attention",
19
- "full_attention",
20
- "sliding_attention",
21
- "full_attention",
22
- "sliding_attention",
23
- "full_attention",
24
- "sliding_attention",
25
- "full_attention",
26
- "sliding_attention",
27
- "full_attention",
28
- "sliding_attention",
29
- "full_attention",
30
- "sliding_attention",
31
- "full_attention",
32
- "sliding_attention",
33
- "full_attention",
34
- "sliding_attention",
35
- "full_attention",
36
- "sliding_attention",
37
- "full_attention",
38
- "sliding_attention",
39
- "full_attention",
40
- "sliding_attention",
41
- "full_attention"
42
- ],
43
- "max_position_embeddings": 131072,
44
- "model_name": "unsloth/gpt-oss-20b-unsloth-bnb-4bit",
45
- "model_type": "gpt_oss",
46
- "num_attention_heads": 64,
47
- "num_experts_per_tok": 4,
48
- "num_hidden_layers": 24,
49
- "num_key_value_heads": 8,
50
- "num_local_experts": 32,
51
- "output_router_logits": false,
52
- "pad_token_id": 200017,
53
- "rms_norm_eps": 1e-05,
54
- "rope_scaling": {
55
- "beta_fast": 32.0,
56
- "beta_slow": 1.0,
57
- "factor": 32.0,
58
- "original_max_position_embeddings": 4096,
59
- "rope_type": "yarn",
60
- "truncate": false
61
- },
62
- "rope_theta": 150000,
63
- "router_aux_loss_coef": 0.9,
64
- "sliding_window": 128,
65
- "swiglu_limit": 7.0,
66
- "tie_word_embeddings": false,
67
- "unsloth_version": "2026.3.3",
68
- "use_cache": true,
69
- "vocab_size": 201088
70
- }
 
1
  {
2
+ "architectures": [
3
+ "GptOssForCausalLM"
4
+ ],
5
+ "attention_bias": true,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 199998,
8
+ "torch_dtype": "bfloat16",
9
+ "eos_token_id": 200002,
10
+ "experts_per_token": 4,
11
+ "head_dim": 64,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 2880,
14
+ "initial_context_length": 4096,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 2880,
17
+ "layer_types": [
18
+ "sliding_attention",
19
+ "full_attention",
20
+ "sliding_attention",
21
+ "full_attention",
22
+ "sliding_attention",
23
+ "full_attention",
24
+ "sliding_attention",
25
+ "full_attention",
26
+ "sliding_attention",
27
+ "full_attention",
28
+ "sliding_attention",
29
+ "full_attention",
30
+ "sliding_attention",
31
+ "full_attention",
32
+ "sliding_attention",
33
+ "full_attention",
34
+ "sliding_attention",
35
+ "full_attention",
36
+ "sliding_attention",
37
+ "full_attention",
38
+ "sliding_attention",
39
+ "full_attention",
40
+ "sliding_attention",
41
+ "full_attention"
42
+ ],
43
+ "max_position_embeddings": 131072,
44
+ "model_name": "unsloth/gpt-oss-20b-unsloth-bnb-4bit",
45
+ "model_type": "gpt_oss",
46
+ "num_attention_heads": 64,
47
+ "num_experts_per_tok": 4,
48
+ "num_hidden_layers": 24,
49
+ "num_key_value_heads": 8,
50
+ "num_local_experts": 32,
51
+ "output_router_logits": false,
52
+ "pad_token_id": 200017,
53
+ "quantization_config": {
54
+ "bnb_4bit_compute_dtype": "bfloat16",
55
+ "bnb_4bit_quant_type": "nf4",
56
+ "bnb_4bit_use_double_quant": true,
57
+ "llm_int8_enable_fp32_cpu_offload": false,
58
+ "llm_int8_has_fp16_weight": false,
59
+ "llm_int8_skip_modules": [
60
+ "lm_head"
61
  ],
62
+ "llm_int8_threshold": 6.0,
63
+ "load_in_4bit": true,
64
+ "load_in_8bit": false,
65
+ "quant_method": "bitsandbytes"
66
+ },
67
+ "rms_norm_eps": 1e-05,
68
+ "rope_scaling": {
69
+ "beta_fast": 32.0,
70
+ "beta_slow": 1.0,
71
+ "factor": 32.0,
72
+ "original_max_position_embeddings": 4096,
73
+ "rope_type": "yarn",
74
+ "truncate": false
75
+ },
76
+ "rope_theta": 150000,
77
+ "router_aux_loss_coef": 0.9,
78
+ "sliding_window": 128,
79
+ "swiglu_limit": 7.0,
80
+ "tie_word_embeddings": false,
81
+ "transformers_version": "4.56.2",
82
+ "unsloth_version": "2026.3.3",
83
+ "use_cache": true,
84
+ "vocab_size": 201088
85
+ }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
generation_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 199998,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 200002,
6
+ 199999
7
+ ],
8
+ "max_length": 131072,
9
+ "pad_token_id": 200017,
10
+ "transformers_version": "4.56.2"
11
+ }
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af455ffddd233dda3590b89dcd6fbcc2b55726acac5707f1751ce51573d28ced
3
+ size 5000435357
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:470d73f5e1f7e1ca091348de0991c9abf2ad8e798d6f30242ea00c3ba3b53678
3
+ size 5000706475
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8010494f3dd9672765080e42018a41bc6e757bec1855021eab11e1820d30caad
3
+ size 2524608106
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff