nm-autobot commited on
Commit
d29337c
·
verified ·
1 Parent(s): 013d775

Upload folder using huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +3 -3
config.json CHANGED
@@ -17,7 +17,7 @@
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,
@@ -30,7 +30,7 @@
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",
@@ -50,7 +50,7 @@
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,
 
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,
 
30
  "dynamic": false,
31
  "group_size": 16,
32
  "num_bits": 4,
33
+ "observer": null,
34
  "observer_kwargs": {},
35
  "scale_dtype": "torch.float8_e4m3fn",
36
  "strategy": "tensor_group",
 
50
  "quant_method": "compressed-tensors",
51
  "quantization_status": "compressed",
52
  "transform_config": {},
53
+ "version": "0.19.1a20261003"
54
  },
55
  "rms_norm_eps": 1e-05,
56
  "rope_scaling": null,