Qwen2-57B-A14B-Instruct

Home/Models/Qwen/Qwen2-57B-A14B-Instruct

Qwen/Qwen2-57B-A14B-Instruct

kind: original
quant: bf16 (16-bit)
family: qwen
modality: text-generation
root model: Qwen/Qwen2-57B-A14B
provider: Qwen
license: apache-2.0
language: en
context: 32768,
library: —
vRAM (GB): —
size on disk: —
downloads: 16,868
likes: 83
created: —
last modified: 2024-08-21T10:24:04+00:00 (2024)
huggingface: https://huggingface.co/Qwen/Qwen2-57B-A14B-Instruct
tags: transformerssafetensorsqwen2_moetext-generationchatconversationalenarxiv:2309.00071base_model:Qwen/Qwen2-57B-A14Bbase_model:finetune:Qwen/Qwen2-57B-A14Blicense:apache-2.0endpoints_compatibleregion:us

benchmark	category	score	context	conf	evidence
C-Eval	knowledge	76.7	C-Eval	medium	`\| C-Eval \| - \| - \| 76.7 \| 80.5 \|`
MMLU	knowledge	71.4	MMLU	medium	`\| MMLU \| 71.4 \| 76.8 \| 74.8 \| 75.4 \|`
GSM8K	math	65.7	GSM8K	medium	`\| GSM8K \| 65.7 \| 90.2 \| 83.6 \| 79.6 \|`
MBPP	coding	59.5	MBPP	medium	`\| MBPP \| 59.5 \| 74.6 \| 67.9 \| 70.9 \|`
EvalPlus	coding	48.5	EvalPlus	medium	`\| EvalPlus \| 48.5 \| - \| 63.6 \| 71.6 \|`
HumanEval	coding	45.1	HumanEval	medium	`\| HumanEval \| 45.1 \| 75.2 \| 68.3 \| 79.9 \|`
MMLU-Pro	knowledge	43.3	MMLU-Pro	medium	`\| MMLU-Pro \| 43.3 \| 52.3 \| 46.4 \| 52.8 \|`
GPQA	reasoning	30.8	GPQA	medium	`\| GPQA \| - \| - \| 30.8 \| 34.3 \|`
MATH	math	30.7	MATH	medium	`\| MATH \| 30.7 \| 50.1 \| 42.4 \| 49.1 \|`
LiveCodeBench	coding	12.3	LiveCodeBench	medium	`\| LiveCodeBench \| 12.3 \| - \| 15.2 \| 25.5 \|`
MT-Bench	chat	8.3	MT-Bench	medium	`\| MT-Bench \| 8.30 \| 8.50 \| 8.30 \| 8.55 \|`

none scraped.