galqiwi commited on
Commit
4cea3cc
1 Parent(s): 1798944

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -1
README.md CHANGED
@@ -12,11 +12,12 @@ base_model: meta-llama/Meta-Llama-3.1-70B-Instruct
12
  ---
13
 
14
  Official [AQLM](https://arxiv.org/abs/2401.06118) quantization of [meta-llama/Meta-Llama-3.1-70B-Instruct
15
- ](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) finetuned with [PV-Tuning](https://arxiv.org/abs/2405.14852).
16
 
17
  For this quantization, we used 1 codebook of 16 bits and groupsize of 8.
18
 
19
  Results:
20
  | Model | Quantization | MMLU (5-shot) | ArcC | ArcE | Hellaswag | PiQA | Winogrande | Model size, Gb |
21
  |-------|--------------|---------------|--------|--------|-----------|--------|------------|----------------|
 
22
  | | 1x16g8 | 0.7814 | 0.5478 | 0.8270 | 0.6284 | 0.8036 | 0.7814 | 21.9 |
 
12
  ---
13
 
14
  Official [AQLM](https://arxiv.org/abs/2401.06118) quantization of [meta-llama/Meta-Llama-3.1-70B-Instruct
15
+ ](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) finetuned with [PV-Tuning](https://arxiv.org/abs/2405.14852).
16
 
17
  For this quantization, we used 1 codebook of 16 bits and groupsize of 8.
18
 
19
  Results:
20
  | Model | Quantization | MMLU (5-shot) | ArcC | ArcE | Hellaswag | PiQA | Winogrande | Model size, Gb |
21
  |-------|--------------|---------------|--------|--------|-----------|--------|------------|----------------|
22
+ | | fp16 | 0.8213 | 0.6246 | 0.8683 | 0.6516 | 0.8313 | 0.7908 | 141 |
23
  | | 1x16g8 | 0.7814 | 0.5478 | 0.8270 | 0.6284 | 0.8036 | 0.7814 | 21.9 |