CoolFace
Modelpublic

LeiGong/vlm_3d_tracking_annotator_opensource_checkpoints

sourceHugging Faceotherupdated 1d agoView on Hugging Face
0likes
Model Card

vlm3dtrackingannotatoropensource_checkpoints

Mirror of the open-source vision-language checkpoints evaluated in the open_source_vlm_test pipeline of https://github.com/XuweiyiChen/vlm3dtracking_annotator (branch lei/open-source-vlm-eval). Each folder is a byte-for-byte copy of one upstream Hugging Face snapshot at the commit we evaluated (or, for folders marked planned, the commit we intend to evaluate), including its LICENSE / terms files; the original licences apply unchanged. Upstream repos remain the canonical source; this mirror exists so the evaluated revisions stay available.

folderupstream repoupstream commitsizestatus
Qwen--Qwen3.5-0.8B/Qwen/Qwen3.5-0.8B2fc06364715b1.6 GBevaluated
Qwen--Qwen3.5-27B-FP8/Qwen/Qwen3.5-27B-FP897f5941bf61728.8 GBevaluated
Qwen--Qwen3.5-2B/Qwen/Qwen3.5-2B15852e8c16364.3 GBevaluated
Qwen--Qwen3.5-4B/Qwen/Qwen3.5-4B851bf6e806ef8.7 GBevaluated
Qwen--Qwen3.5-9B/Qwen/Qwen3.5-9Bc2022362357618.0 GBevaluated
Qwen--Qwen3.8-27B-FP8/Qwen/Qwen3.8-27B-FP8017b9c7af6b528.8 GBevaluated
RedHatAI--gemma-4-31B-it-FP8-dynamic/RedHatAI/gemma-4-31B-it-FP8-dynamicd4ab4f579dd331.0 GBevaluated
allenai--Molmo2-VideoPoint-4B/allenai/Molmo2-VideoPoint-4B7deb981c9a8a18.1 GBplanned, not yet evaluated: Molmo2 architecture is served by vLLM 0.29 (trustremotecode); fits one 48 GB GPU
allenai--MolmoPoint-8B/allenai/MolmoPoint-8B188130f961c832.4 GBevaluated
meta-models--Muse-Glimmer-30B/meta-models/Muse-Glimmer-30Ba4e59da52a7b55.5 GBplanned, not yet evaluated: image-text-to-text, MuseGlimmerForConditionalGeneration is served by vLLM 0.29; 60 GB of bf16 weights need two 48 GB GPUs
zai-org--GLM-5.3-Flash/zai-org/GLM-5.3-Flasheb9eb208eb0d305.8 GBplanned, not yet evaluated: needs a vLLM release with the Glm5Next architecture and >= 4 x 80 GB Hopper GPUs (306 GB of FP8 weights)

Load one folder with huggingface_hub.snapshot_download(repo_id, allow_patterns=['<folder>/*']) or point vllm serve / transformers at the local sub-directory.

See MIRROR_MANIFEST.json for the full record (file counts, mirror commits).