nm-autobot commited on
Commit
ebff91b
·
verified ·
1 Parent(s): 844fbb9

Upload folder using huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +8 -4
config.json CHANGED
@@ -17,7 +17,7 @@
17
  "pretraining_tp": 1,
18
  "quantization_config": {
19
  "config_groups": {
20
- "NVFP4A16": {
21
  "format": "nvfp4-pack-quantized",
22
  "input_activations": null,
23
  "output_activations": null,
@@ -25,13 +25,18 @@
25
  "Linear"
26
  ],
27
  "weights": {
 
 
28
  "dynamic": false,
29
  "group_size": 16,
30
  "num_bits": 4,
 
 
31
  "scale_dtype": "torch.float8_e4m3fn",
32
  "strategy": "tensor_group",
33
  "symmetric": true,
34
- "type": "float"
 
35
  }
36
  }
37
  },
@@ -42,11 +47,10 @@
42
  "re:.*embed_tokens.*"
43
  ],
44
  "kv_cache_scheme": null,
45
- "moe_sparsity_config": null,
46
  "quant_method": "compressed-tensors",
47
  "quantization_status": "compressed",
48
  "transform_config": {},
49
- "version": "0.20.2a20261028"
50
  },
51
  "rms_norm_eps": 1e-05,
52
  "rope_scaling": null,
 
17
  "pretraining_tp": 1,
18
  "quantization_config": {
19
  "config_groups": {
20
+ "config_group_0": {
21
  "format": "nvfp4-pack-quantized",
22
  "input_activations": null,
23
  "output_activations": null,
 
25
  "Linear"
26
  ],
27
  "weights": {
28
+ "actorder": null,
29
+ "block_structure": null,
30
  "dynamic": false,
31
  "group_size": 16,
32
  "num_bits": 4,
33
+ "observer": "memoryless_minmax",
34
+ "observer_kwargs": {},
35
  "scale_dtype": "torch.float8_e4m3fn",
36
  "strategy": "tensor_group",
37
  "symmetric": true,
38
+ "type": "float",
39
+ "zp_dtype": null
40
  }
41
  }
42
  },
 
47
  "re:.*embed_tokens.*"
48
  ],
49
  "kv_cache_scheme": null,
 
50
  "quant_method": "compressed-tensors",
51
  "quantization_status": "compressed",
52
  "transform_config": {},
53
+ "version": "0.17.1"
54
  },
55
  "rms_norm_eps": 1e-05,
56
  "rope_scaling": null,