nm-autobot commited on
Commit
a69eaa4
·
verified ·
1 Parent(s): 0149e06

Upload folder using huggingface_hub

Browse files
Files changed (4) hide show
  1. config.json +4 -13
  2. generation_config.json +1 -1
  3. model.safetensors +1 -1
  4. recipe.yaml +4 -26
config.json CHANGED
@@ -20,36 +20,26 @@
20
  "group_0": {
21
  "format": "int-quantized",
22
  "input_activations": {
23
- "actorder": null,
24
- "block_structure": null,
25
  "dynamic": false,
26
  "group_size": null,
27
  "num_bits": 8,
28
  "observer": "static_minmax",
29
- "observer_kwargs": {},
30
- "scale_dtype": null,
31
  "strategy": "tensor",
32
  "symmetric": false,
33
- "type": "int",
34
- "zp_dtype": "torch.int8"
35
  },
36
  "output_activations": null,
37
  "targets": [
38
  "Linear"
39
  ],
40
  "weights": {
41
- "actorder": null,
42
- "block_structure": null,
43
  "dynamic": false,
44
  "group_size": null,
45
  "num_bits": 8,
46
  "observer": "memoryless_minmax",
47
- "observer_kwargs": {},
48
- "scale_dtype": null,
49
  "strategy": "channel",
50
  "symmetric": true,
51
- "type": "int",
52
- "zp_dtype": null
53
  }
54
  }
55
  },
@@ -59,11 +49,12 @@
59
  "lm_head"
60
  ],
61
  "kv_cache_scheme": null,
 
62
  "quant_method": "compressed-tensors",
63
  "quantization_status": "compressed",
64
  "sparsity_config": {},
65
  "transform_config": {},
66
- "version": "0.19.0"
67
  },
68
  "rms_norm_eps": 1e-05,
69
  "rope_scaling": null,
 
20
  "group_0": {
21
  "format": "int-quantized",
22
  "input_activations": {
 
 
23
  "dynamic": false,
24
  "group_size": null,
25
  "num_bits": 8,
26
  "observer": "static_minmax",
 
 
27
  "strategy": "tensor",
28
  "symmetric": false,
29
+ "type": "int"
 
30
  },
31
  "output_activations": null,
32
  "targets": [
33
  "Linear"
34
  ],
35
  "weights": {
 
 
36
  "dynamic": false,
37
  "group_size": null,
38
  "num_bits": 8,
39
  "observer": "memoryless_minmax",
 
 
40
  "strategy": "channel",
41
  "symmetric": true,
42
+ "type": "int"
 
43
  }
44
  }
45
  },
 
49
  "lm_head"
50
  ],
51
  "kv_cache_scheme": null,
52
+ "moe_sparsity_config": null,
53
  "quant_method": "compressed-tensors",
54
  "quantization_status": "compressed",
55
  "sparsity_config": {},
56
  "transform_config": {},
57
+ "version": "0.19.1a20261022"
58
  },
59
  "rms_norm_eps": 1e-05,
60
  "rope_scaling": null,
generation_config.json CHANGED
@@ -3,5 +3,5 @@
3
  "eos_token_id": 2,
4
  "max_length": 2048,
5
  "pad_token_id": 0,
6
- "transformers_version": "5.17.0"
7
  }
 
3
  "eos_token_id": 2,
4
  "max_length": 2048,
5
  "pad_token_id": 0,
6
+ "transformers_version": "5.20.2"
7
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:800c246431bd6730cce6b001abff906b786b1aae94c26690cae405d8b210172b
3
  size 1232076790
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c24752ccaf572e1126fa09699231943573db1ff782403eb14d1b65bbcb62841
3
  size 1232076790
recipe.yaml CHANGED
@@ -16,32 +16,10 @@ quant_stage:
16
  config_groups:
17
  group_0:
18
  targets: [Linear]
19
- weights:
20
- num_bits: 8
21
- type: int
22
- symmetric: true
23
- group_size: null
24
- strategy: channel
25
- block_structure: null
26
- dynamic: false
27
- actorder: null
28
- scale_dtype: null
29
- zp_dtype: null
30
- observer: memoryless_minmax
31
- observer_kwargs: {}
32
- input_activations:
33
- num_bits: 8
34
- type: int
35
- symmetric: false
36
- group_size: null
37
- strategy: tensor
38
- block_structure: null
39
- dynamic: false
40
- actorder: null
41
- scale_dtype: null
42
- zp_dtype: torch.int8
43
- observer: static_minmax
44
- observer_kwargs: {}
45
  output_activations: null
46
  format: null
47
  targets: [Linear]
 
16
  config_groups:
17
  group_0:
18
  targets: [Linear]
19
+ weights: {num_bits: 8, type: int, symmetric: true, group_size: null, strategy: channel,
20
+ dynamic: false, observer: memoryless_minmax}
21
+ input_activations: {num_bits: 8, type: int, symmetric: false, group_size: null,
22
+ strategy: tensor, dynamic: false, observer: static_minmax}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  output_activations: null
24
  format: null
25
  targets: [Linear]