The table will scroll to the left
Task name | Result | Metric |
---|---|---|
LCS | 0.07 | Accuracy |
RCB | 0.518 / 0.441 | Avg. F1 / Accuracy |
USE | 0.334 | Grade Norm |
RWSD | 0.512 | Accuracy |
PARus | 0.842 | Accuracy |
ruTiE | 0.758 | Accuracy |
MultiQ | 0.393 / 0.187 | F1-score/EM |
CheGeKa | 0.318 / 0.252 | F1 / EM |
ruModAr | 0.87 | EM |
MaMuRAMu | 0.741 | Accuracy |
ruMultiAr | 0.272 | EM |
ruCodeEval | 0.041 / 0.054 / 0.061 | pass@k |
MathLogicQA | 0.455 | Accuracy |
ruWorldTree | 0.901 / 0.901 | Avg. F1 / Accuracy |
ruOpenBookQA | 0.833 / 0.833 | Avg. F1 / Accuracy |
The table will scroll to the left
Task name | Result | Metric | ||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
BPS | 0.921 | Accuracy | ||||||||||||||||||||||||
ruMMLU | 0.587 | Accuracy | ||||||||||||||||||||||||
SimpleAr | 0.923 | EM | ||||||||||||||||||||||||
ruHumanEval | 0.037 / 0.04 / 0.043 | pass@k | ||||||||||||||||||||||||
ruHHH |
0.73
|
Accuracy | ||||||||||||||||||||||||
ruHateSpeech |
0.777
|
Accuracy | ||||||||||||||||||||||||
ruDetox |
|
Overall average score (J) Assessment of the preservation of meaning (SIM) Assessment of naturalness (FL) Style Transfer Accuracy (STA) |
||||||||||||||||||||||||
ruEthics |
Table results:
[[0.299, 0.337
, 0.351, 0.251
, 0.249], |
5 MCC |
GIGACHAT
GigaChat-20B-A3B
20.0B
Opened
SFT
MoE
GigaChat-20B-A3B is a Large Language Model (LLM) that was fine-tuned on instruction corpus and has context length of 32k tokens. GigaChat-20B-A3B is Mixture of Experts model and has 3.3B active parameters. The model is available at https://huggingface.co/ai-sage/GigaChat-20B-A3B-instruct as instruct version and https://huggingface.co/ai-sage/GigaChat-20B-A3B-base as base version
-
-
Open-source model by Sber
MERA version:
v.1.2.0
The version of the codebase:
db539c9
Torch Version:
2.4.0
CUDA version:
12.1
Transformers version:
4.46.0.dev0
The number of GPUs and their type:
5 x NVIDIA H100 80GB HBM3
Butch:
1
Seed:
1234
Architecture:
gigachat_llms
Chat template:
Yes
Special tokens:
Yes
Multi-Turn:
1
Generation Parameters:
simplear - do_sample=false;until=["\n"];
chegeka - do_sample=false;until=["\n"];
rudetox - do_sample=false;until=["\n"];
rumultiar - do_sample=false;until=["\n"];
use - do_sample=false;until=["\n","."];
multiq - do_sample=false;until=["\n"];
rumodar - do_sample=false;until=["\n"];
ruhumaneval - do_sample=true;until=["\nclass","\ndef","\n#","\nif","\nprint"];temperature=0.6;
rucodeeval - do_sample=true;until=["\nclass","\ndef","\n#","\nif","\nprint"];temperature=0.6;
Description of the template:
{% if messages[0]['role'] == 'system' -%}\n {%- set loop_messages = messages[1:] -%}\n {%- set system_message = bos_token + messages[0]['content'] + additional_special_tokens[1] -%}\n{%- else -%}\n {%- set loop_messages = messages -%}\n {%- set system_message = bos_token + '' -%}\n{%- endif -%}\n{%- for message in loop_messages %}\n {% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n {{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}\n {% endif %}\n \n {%- if loop.index0 == 0 -%}\n {{ system_message -}}\n {%- endif -%}\n {%- if message['role'] == 'user' -%}\n {{ message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1] -}}\n {{ 'available functions' + additional_special_tokens[0] + additional_special_tokens[2] + additional_special_tokens[3] + additional_special_tokens[1] -}}\n {%- endif -%}\n {%- if message['role'] == 'assistant' -%}\n {{ message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1] -}}\n {%- endif -%}\n {%- if loop.last and add_generation_prompt -%}\n {{ 'assistant' + additional_special_tokens[0] -}}\n {%- endif -%}\n{%- endfor %}