Hugging Face's logo

HeyMiku
/
Intent-Classifier-4B

Text Generation
MLX
Safetensors
English
qwen3_5
conversational
Model card Files
xet
 Community
HeyMiku commited on
Commit
7588fa9
·
verified ·
1 Parent(s): 9c90cb4

Replace MLX 4-bit with BF16 merged model (SFT best@6k, val loss 0.015). Compatible with vLLM/transformers.

Browse files
Files changed (4) hide show
  1. config.json +81 -96
  2. generation_config.json +6 -0
  3. model.safetensors +2 -2
  4. tokenizer_config.json +0 -2
config.json CHANGED
@@ -1,98 +1,83 @@
1
  {
2
- "architectures": [
3
- "Qwen3_5ForConditionalGeneration"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  ],
5
- "image_token_id": 248056,
6
- "model_type": "qwen3_5",
7
- "quantization": {
8
- "group_size": 64,
9
- "bits": 4,
10
- "mode": "affine"
11
- },
12
- "quantization_config": {
13
- "group_size": 64,
14
- "bits": 4,
15
- "mode": "affine"
16
- },
17
- "text_config": {
18
- "attention_bias": false,
19
- "attention_dropout": 0.0,
20
- "attn_output_gate": true,
21
- "dtype": "bfloat16",
22
- "eos_token_id": 248044,
23
- "full_attention_interval": 4,
24
- "head_dim": 256,
25
- "hidden_act": "silu",
26
- "hidden_size": 2560,
27
- "initializer_range": 0.02,
28
- "intermediate_size": 9216,
29
- "layer_types": [
30
- "linear_attention",
31
- "linear_attention",
32
- "linear_attention",
33
- "full_attention",
34
- "linear_attention",
35
- "linear_attention",
36
- "linear_attention",
37
- "full_attention",
38
- "linear_attention",
39
- "linear_attention",
40
- "linear_attention",
41
- "full_attention",
42
- "linear_attention",
43
- "linear_attention",
44
- "linear_attention",
45
- "full_attention",
46
- "linear_attention",
47
- "linear_attention",
48
- "linear_attention",
49
- "full_attention",
50
- "linear_attention",
51
- "linear_attention",
52
- "linear_attention",
53
- "full_attention",
54
- "linear_attention",
55
- "linear_attention",
56
- "linear_attention",
57
- "full_attention",
58
- "linear_attention",
59
- "linear_attention",
60
- "linear_attention",
61
- "full_attention"
62
- ],
63
- "linear_conv_kernel_dim": 4,
64
- "linear_key_head_dim": 128,
65
- "linear_num_key_heads": 16,
66
- "linear_num_value_heads": 32,
67
- "linear_value_head_dim": 128,
68
- "max_position_embeddings": 262144,
69
- "mlp_only_layers": [],
70
- "model_type": "qwen3_5_text",
71
- "mtp_num_hidden_layers": 1,
72
- "mtp_use_dedicated_embeddings": false,
73
- "num_attention_heads": 16,
74
- "num_hidden_layers": 32,
75
- "num_key_value_heads": 4,
76
- "rms_norm_eps": 1e-06,
77
- "tie_word_embeddings": true,
78
- "use_cache": true,
79
- "vocab_size": 248320,
80
- "mamba_ssm_dtype": "float32",
81
- "rope_parameters": {
82
- "mrope_interleaved": true,
83
- "mrope_section": [
84
- 11,
85
- 11,
86
- 10
87
- ],
88
- "rope_theta": 10000000,
89
- "partial_rotary_factor": 0.25,
90
- "type": "default"
91
- }
92
- },
93
- "tie_word_embeddings": true,
94
- "transformers_version": "4.57.0.dev0",
95
- "video_token_id": 248057,
96
- "vision_end_token_id": 248054,
97
- "vision_start_token_id": 248053
98
- }
 
1
  {
2
+ "architectures": [
3
+ "Qwen3_5ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "attn_output_gate": true,
8
+ "bos_token_id": null,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": 248044,
11
+ "full_attention_interval": 4,
12
+ "head_dim": 256,
13
+ "hidden_act": "silu",
14
+ "hidden_size": 2560,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 9216,
17
+ "layer_types": [
18
+ "linear_attention",
19
+ "linear_attention",
20
+ "linear_attention",
21
+ "full_attention",
22
+ "linear_attention",
23
+ "linear_attention",
24
+ "linear_attention",
25
+ "full_attention",
26
+ "linear_attention",
27
+ "linear_attention",
28
+ "linear_attention",
29
+ "full_attention",
30
+ "linear_attention",
31
+ "linear_attention",
32
+ "linear_attention",
33
+ "full_attention",
34
+ "linear_attention",
35
+ "linear_attention",
36
+ "linear_attention",
37
+ "full_attention",
38
+ "linear_attention",
39
+ "linear_attention",
40
+ "linear_attention",
41
+ "full_attention",
42
+ "linear_attention",
43
+ "linear_attention",
44
+ "linear_attention",
45
+ "full_attention",
46
+ "linear_attention",
47
+ "linear_attention",
48
+ "linear_attention",
49
+ "full_attention"
50
+ ],
51
+ "linear_conv_kernel_dim": 4,
52
+ "linear_key_head_dim": 128,
53
+ "linear_num_key_heads": 16,
54
+ "linear_num_value_heads": 32,
55
+ "linear_value_head_dim": 128,
56
+ "mamba_ssm_dtype": "float32",
57
+ "max_position_embeddings": 262144,
58
+ "mlp_only_layers": [],
59
+ "model_type": "qwen3_5_text",
60
+ "mtp_num_hidden_layers": 1,
61
+ "mtp_use_dedicated_embeddings": false,
62
+ "num_attention_heads": 16,
63
+ "num_hidden_layers": 32,
64
+ "num_key_value_heads": 4,
65
+ "pad_token_id": null,
66
+ "partial_rotary_factor": 0.25,
67
+ "rms_norm_eps": 1e-06,
68
+ "rope_parameters": {
69
+ "mrope_interleaved": true,
70
+ "mrope_section": [
71
+ 11,
72
+ 11,
73
+ 10
74
  ],
75
+ "partial_rotary_factor": 0.25,
76
+ "rope_theta": 10000000,
77
+ "rope_type": "default"
78
+ },
79
+ "tie_word_embeddings": true,
80
+ "transformers_version": "5.5.0",
81
+ "use_cache": true,
82
+ "vocab_size": 248320
83
+ }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "eos_token_id": 248044,
4
+ "transformers_version": "5.5.0",
5
+ "use_cache": true
6
+ }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:899c5685debb3a7529a9e36ee351abd078843ff615bab763ab26dad8a20a99fe
3
- size 2367238661
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b07090f8b28aee43f9f1d5882e2bfd42053de88a6269b20b025f596519dcbd0d
3
+ size 8411558400
tokenizer_config.json CHANGED
@@ -22,10 +22,8 @@
22
  },
23
  "pad_token": "<|endoftext|>",
24
  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
25
- "processor_class": "Qwen3VLProcessor",
26
  "split_special_tokens": false,
27
  "tokenizer_class": "TokenizersBackend",
28
- "tool_parser_type": "qwen3_coder",
29
  "unk_token": null,
30
  "video_token": "<|video_pad|>",
31
  "vision_bos_token": "<|vision_start|>",
 
22
  },
23
  "pad_token": "<|endoftext|>",
24
  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
 
25
  "split_special_tokens": false,
26
  "tokenizer_class": "TokenizersBackend",
 
27
  "unk_token": null,
28
  "video_token": "<|video_pad|>",
29
  "vision_bos_token": "<|vision_start|>",