Hugging Face's logo

Intel
/
Llama-3.3-70B-Instruct-AutoRound-Recipe

Model card Files
xet
 Community
INC4AI commited on
Commit
67d9043
·
verified ·
1 Parent(s): 791d781

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -6
README.md CHANGED
@@ -15,12 +15,13 @@ The step-by-step README of quantization and evaluation can be found in [Intel Ne
15
 
16
  ## Evaluate Results
17
 
18
- | Task | backend | BF16 | MXFP8 | MXFP4(mixed) |
19
- |:-----------:|:-------:|:------:|:------:|:------------:|
20
- | hellaswag | vllm | 0.6661 | 0.6660 | 0.6567 |
21
- | piqa | vllm | 0.8357 | 0.8324 | 0.8303 |
22
- | mmlu_llama | vllm | 0.8336 | 0.8326 | 0.8291 |
23
- | gsm8k_llama | vllm | 0.9492 | 0.9439 | 0.9416 |
 
24
 
25
 
26
  ## Ethical Considerations and Limitations
 
15
 
16
  ## Evaluate Results
17
 
18
+ | Task | backend | BF16 | MXFP8 | MXFP4(mixed) |
19
+ |:-------------------:|:-------:|:------:|:------:|:------------:|
20
+ | hellaswag | vllm | 0.6661 | 0.6660 | 0.6567 |
21
+ | piqa | vllm | 0.8357 | 0.8324 | 0.8303 |
22
+ | mmlu_llama | vllm | 0.8336 | 0.8326 | 0.8291 |
23
+ | gsm8k_llama(strict) | vllm | 0.9492 | 0.9439 | 0.9416 |
24
+ | average | vllm | 0.8211 | 0.8187 | 0.8144 |
25
 
26
 
27
  ## Ethical Considerations and Limitations