Hugging Face's logo

noneUsername
/
Forgotten-Safeword-24B-V3.0-W8A8

Safetensors
mistral
8-bit precision
compressed-tensors
Model card Files Community
noneUsername commited on
Commit
14ff4aa
·
verified ·
1 Parent(s): 7632de8

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +68 -0
README.md ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model:
3
+ - ReadyArt/Forgotten-Safeword-24B-V3.0
4
+ ---
5
+ vllm (pretrained=/root/autodl-tmp/Forgotten-Safeword-24B-V3.0,add_bos_token=true,max_model_len=4096,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
6
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
7
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
8
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ | 0.92|± |0.0172|
9
+ | | |strict-match | 5|exact_match|↑ | 0.92|± |0.0172|
10
+
11
+ vllm (pretrained=/root/autodl-tmp/Forgotten-Safeword-24B-V3.0,add_bos_token=true,max_model_len=4096,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto
12
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
13
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
14
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.912|± |0.0127|
15
+ | | |strict-match | 5|exact_match|↑ |0.904|± |0.0132|
16
+
17
+ vllm (pretrained=/root/autodl-tmp/Forgotten-Safeword-24B-V3.0,add_bos_token=true,max_model_len=4096,dtype=bfloat16,max_num_seqs=3), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1
18
+ | Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
19
+ |------------------|------:|------|------|------|---|-----:|---|-----:|
20
+ |mmlu | 2|none | |acc |↑ |0.7953|± |0.0131|
21
+ | - humanities | 2|none | |acc |↑ |0.8000|± |0.0270|
22
+ | - other | 2|none | |acc |↑ |0.8051|± |0.0272|
23
+ | - social sciences| 2|none | |acc |↑ |0.8667|± |0.0244|
24
+ | - stem | 2|none | |acc |↑ |0.7404|± |0.0248|
25
+
26
+
27
+ vllm (pretrained=/root/autodl-tmp/70-128-df10,add_bos_token=true,max_model_len=4096,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
28
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
29
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
30
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.888|± |0.0200|
31
+ | | |strict-match | 5|exact_match|↑ |0.884|± |0.0203|
32
+
33
+ vllm (pretrained=/root/autodl-tmp/70-128-df10,add_bos_token=true,max_model_len=4096,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto
34
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
35
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
36
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.904|± |0.0132|
37
+ | | |strict-match | 5|exact_match|↑ |0.894|± |0.0138|
38
+
39
+ vllm (pretrained=/root/autodl-tmp/70-128-df10,add_bos_token=true,max_model_len=4096,dtype=bfloat16,max_num_seqs=3), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1
40
+ | Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
41
+ |------------------|------:|------|------|------|---|-----:|---|-----:|
42
+ |mmlu | 2|none | |acc |↑ |0.7860|± |0.0131|
43
+ | - humanities | 2|none | |acc |↑ |0.8051|± |0.0252|
44
+ | - other | 2|none | |acc |↑ |0.7846|± |0.0276|
45
+ | - social sciences| 2|none | |acc |↑ |0.8667|± |0.0240|
46
+ | - stem | 2|none | |acc |↑ |0.7228|± |0.0255|
47
+
48
+
49
+ vllm (pretrained=/root/autodl-tmp/86-512,add_bos_token=true,max_model_len=4096,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto
50
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
51
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
52
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ | 0.94|± |0.0151|
53
+ | | |strict-match | 5|exact_match|↑ | 0.94|± |0.0151|
54
+
55
+ vllm (pretrained=/root/autodl-tmp/86-512,add_bos_token=true,max_model_len=4096,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto
56
+ |Tasks|Version| Filter |n-shot| Metric | |Value| |Stderr|
57
+ |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
58
+ |gsm8k| 3|flexible-extract| 5|exact_match|↑ |0.920|± |0.0121|
59
+ | | |strict-match | 5|exact_match|↑ |0.916|± |0.0124|
60
+
61
+ vllm (pretrained=/root/autodl-tmp/86-512,add_bos_token=true,max_model_len=4096,dtype=bfloat16,max_num_seqs=3), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: 1
62
+ | Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
63
+ |------------------|------:|------|------|------|---|-----:|---|-----:|
64
+ |mmlu | 2|none | |acc |↑ |0.8012|± |0.0130|
65
+ | - humanities | 2|none | |acc |↑ |0.8000|± |0.0267|
66
+ | - other | 2|none | |acc |↑ |0.8000|± |0.0275|
67
+ | - social sciences| 2|none | |acc |↑ |0.8778|± |0.0234|
68
+ | - stem | 2|none | |acc |↑ |0.7544|± |0.0246|