Skip to page content
SAVRN
Search Plan a project

Open-weight model

visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean

by Zeyuan Yang Miiche/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean

visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean is an open-weight model from Zeyuan Yang. Its published files total 30.1 GB.

Parameters—
Context—
Weights14.6 GB
License—
AccessOpen weights
Monthly Downloads—

Model Card

The publisher has not written a card for this model.

Identity and Version

Repository
Miiche/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean
Publisher
Zeyuan Yang
Task
Not stated by the source
Modality
Other
Library
Not stated by the source
Parameters
Not stated by the source
Languages
Not stated by the source
Revision
ec7d8e83f212ddb4ec99969ba6e50c04c3600488
First published
2026-09-21
Last updated
2026-09-22

Files and Weights

155 files, 30.1 GB in total. The weights are 3 files totalling 14.6 GB in safetensors.

Weights3 files · 14.6 GB
Configuration94 files · 302.1 MB
Tokenizer12 files · 47.6 MB
Other45 files · 15.1 GB
Repository1 file · 2.8 KB
Every file
FileTypeSizeSHA-256
global_step_180_merged/model.safetensorsWeights4.9 GB 8e537a909787
global_step_270_merged/model.safetensorsWeights4.9 GB 4e33d2d7455f
global_step_90_merged/model.safetensorsWeights4.9 GB 3eaa72cb3d93
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/blink_output.jsonConfiguration9.7 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/blink_scores.jsonConfiguration3.2 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/chartqa_output.jsonConfiguration3.2 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/chartqa_scores.jsonConfiguration193 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/hallusionbench_output.jsonConfiguration3.3 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/hallusionbench_scores.jsonConfiguration477 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/logicvista_output.jsonConfiguration4.6 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/logicvista_scores.jsonConfiguration489 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathverse_output.jsonConfiguration25.5 MB a2024241165c
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathverse_scores.jsonConfiguration2.3 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathvision_output.jsonConfiguration4.0 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathvision_scores.jsonConfiguration8.8 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathvista_output.jsonConfiguration6.3 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathvista_scores.jsonConfiguration7.5 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmbench_output.jsonConfiguration14.1 MB 9b95e556acb2
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmbench_scores.jsonConfiguration6.4 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mme_output.jsonConfiguration5.8 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mme_scores.jsonConfiguration2.0 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmmu_output.jsonConfiguration6.1 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmmu_pro_output.jsonConfiguration196.4 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmmu_pro_scores.jsonConfiguration110 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmmu_scores.jsonConfiguration2.6 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmstar_output.jsonConfiguration7.0 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmstar_scores.jsonConfiguration3.1 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmvet_output.jsonConfiguration472.1 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmvet_scores.jsonConfiguration514 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/perceptionbench_output.jsonConfiguration24.0 MB b5d4bdc5b15f
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/perceptionbench_scores.jsonConfiguration1.1 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/realworldqa_output.jsonConfiguration2.2 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/realworldqa_scores.jsonConfiguration95 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/scienceqa_output.jsonConfiguration6.2 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/scienceqa_scores.jsonConfiguration359 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/visnumbench_output.jsonConfiguration11.1 MB 8e70a3a00e1b
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/visnumbench_scores.jsonConfiguration953 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/chartqa_output.jsonConfiguration2.4 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/chartqa_scores.jsonConfiguration191 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/mmvet_output.jsonConfiguration420.9 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/mmvet_scores.jsonConfiguration490 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/realworldqa_output.jsonConfiguration1.6 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/realworldqa_scores.jsonConfiguration95 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/visnumbench_output.jsonConfiguration9.3 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/visnumbench_scores.jsonConfiguration951 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/blink_output.jsonConfiguration10.2 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/blink_scores.jsonConfiguration3.2 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/chartqa_output.jsonConfiguration3.4 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/chartqa_scores.jsonConfiguration193 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/hallusionbench_output.jsonConfiguration3.8 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/hallusionbench_scores.jsonConfiguration493 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/logicvista_output.jsonConfiguration5.0 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/logicvista_scores.jsonConfiguration502 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathverse_output.jsonConfiguration31.2 MB d49652e7420e
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathverse_scores.jsonConfiguration2.3 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathvision_output.jsonConfiguration4.4 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathvision_scores.jsonConfiguration8.8 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathvista_output.jsonConfiguration7.3 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathvista_scores.jsonConfiguration7.5 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmbench_output.jsonConfiguration15.4 MB 1c17b0b8a773
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmbench_scores.jsonConfiguration6.3 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mme_output.jsonConfiguration5.3 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mme_scores.jsonConfiguration2.0 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmmu_output.jsonConfiguration7.2 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmmu_pro_output.jsonConfiguration196.3 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmmu_pro_scores.jsonConfiguration110 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmmu_scores.jsonConfiguration2.7 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmstar_output.jsonConfiguration8.3 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmstar_scores.jsonConfiguration3.1 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmvet_output.jsonConfiguration428.1 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmvet_scores.jsonConfiguration516 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/perceptionbench_output.jsonConfiguration29.8 MB e87c32a31b76
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/perceptionbench_scores.jsonConfiguration1.1 KB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/realworldqa_output.jsonConfiguration2.6 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/realworldqa_scores.jsonConfiguration95 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/scienceqa_output.jsonConfiguration7.5 MB —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/scienceqa_scores.jsonConfiguration359 B —
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/visnumbench_output.jsonConfiguration12.7 MB 5bbeee301c0b
eval_results/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/visnumbench_scores.jsonConfiguration953 B —
global_step_180_merged/added_tokens.jsonConfiguration707 B —
global_step_180_merged/config.jsonConfiguration1.6 KB —
global_step_180_merged/generation_config.jsonConfiguration147 B —
global_step_180_merged/preprocessor_config.jsonConfiguration782 B —
global_step_180_merged/special_tokens_map.jsonConfiguration613 B —
global_step_180_merged/video_preprocessor_config.jsonConfiguration817 B —
global_step_270_merged/added_tokens.jsonConfiguration707 B —
global_step_270_merged/config.jsonConfiguration1.6 KB —
global_step_270_merged/generation_config.jsonConfiguration147 B —
global_step_270_merged/preprocessor_config.jsonConfiguration782 B —
global_step_270_merged/special_tokens_map.jsonConfiguration613 B —
global_step_270_merged/video_preprocessor_config.jsonConfiguration817 B —
global_step_90_merged/added_tokens.jsonConfiguration707 B —
global_step_90_merged/config.jsonConfiguration1.6 KB —
global_step_90_merged/generation_config.jsonConfiguration147 B —
global_step_90_merged/preprocessor_config.jsonConfiguration782 B —
global_step_90_merged/special_tokens_map.jsonConfiguration613 B —
global_step_90_merged/video_preprocessor_config.jsonConfiguration817 B —
global_step_180_merged/chat_template.jinjaOther5.3 KB —
global_step_270_merged/chat_template.jinjaOther5.3 KB —
global_step_90_merged/chat_template.jinjaOther5.3 KB —
logs/eval-train-q3vl2b_ppo1_4bx_c10_Lall.logOther15.1 GB 3f082e75428e
logs/eval-visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm.logOther1.6 MB —
logs/eval-visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm.logOther1.6 MB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/blink.logOther145.4 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/chartqa.logOther105.5 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/hallusionbench.logOther50.5 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/logicvista.logOther31.2 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathverse.logOther196.7 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathvision.logOther28.0 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mathvista.logOther61.6 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmbench.logOther164.3 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mme.logOther108.9 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmmu.logOther84.9 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmmu_pro.logOther116.0 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmstar.logOther73.3 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/mmvet.logOther45.6 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/perceptionbench.logOther183.2 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/realworldqa.logOther59.0 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/scienceqa.logOther86.2 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s180-vllm/visnumbench.logOther88.9 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/chartqa.logOther101.2 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/mmvet.logOther46.2 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/realworldqa.logOther58.1 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s270-vllm/visnumbench.logOther90.9 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/blink.logOther146.0 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/chartqa.logOther108.7 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/hallusionbench.logOther51.1 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/logicvista.logOther32.6 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathverse.logOther197.7 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathvision.logOther25.8 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mathvista.logOther61.6 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmbench.logOther162.8 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mme.logOther110.1 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmmu.logOther81.8 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmmu_pro.logOther113.1 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmstar.logOther75.0 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/mmvet.logOther46.1 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/perceptionbench.logOther184.1 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/realworldqa.logOther61.3 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/scienceqa.logOther88.2 KB —
logs/visualrl-qwen3vl2b-ppo1-4bx-c10-Lall-clean-s90-vllm/visnumbench.logOther85.2 KB —
train_logs/train_metrics_q3vl2b_ppo1_4bx_c10_Lall.jsonlOther544.0 KB —
.gitattributesRepository2.8 KB —
global_step_180_merged/merges.txtTokenizer1.7 MB —
global_step_180_merged/tokenizer.jsonTokenizer11.4 MB aeb13307a71a
global_step_180_merged/tokenizer_config.jsonTokenizer5.5 KB —
global_step_180_merged/vocab.jsonTokenizer2.8 MB —
global_step_270_merged/merges.txtTokenizer1.7 MB —
global_step_270_merged/tokenizer.jsonTokenizer11.4 MB aeb13307a71a
global_step_270_merged/tokenizer_config.jsonTokenizer5.5 KB —
global_step_270_merged/vocab.jsonTokenizer2.8 MB —
global_step_90_merged/merges.txtTokenizer1.7 MB —
global_step_90_merged/tokenizer.jsonTokenizer11.4 MB aeb13307a71a
global_step_90_merged/tokenizer_config.jsonTokenizer5.5 KB —
global_step_90_merged/vocab.jsonTokenizer2.8 MB —

License and Download

License
Not stated by the source
Access
Open weights, no gate
Download size
14.6 GB
Download from Zeyuan Yang

Released by Zeyuan Yang through its official repository on Hugging Face.

Memory Requirements

PrecisionWeights in memory
As published14.6 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.