Describe vision-encoder downstream evaluation checkpoints

README.md

@@ -0,0 +1,51 @@
+---
+tags:
+  - vision-encoder-evaluation
+  - multimodal
+  - visual-tokenizer-benchmark
+  - checkpoints
+---
+
+# Vision Encoder MLLM Evaluation
+
+Downstream evaluation checkpoints for comparing **vision encoders** in multimodal large language models (MLLMs), from the VisualTokenizerBench / RAVEL project.
+
+The evaluation target is the vision encoder. Runs use different language backbones and are organized by backbone, vision encoder, and training stage. These are trained multimodal evaluation artifacts, rather than a standalone language-model release.
+
+**Upload status:** In progress: the expected inventory is available in FILES.json; some checkpoints may not yet be uploaded.
+
+## Contents
+
+| Language-backbone group | Files | Weight files | Size (GiB) |
+| --- | ---: | ---: | ---: |
+| qwen3base | 35 | 8 | 10.46 |
+| smollm2 | 974 | 215 | 285.01 |
+| qwen3 | 978 | 214 | 305.64 |
+| qwen25 | 1173 | 268 | 326.19 |
+
+Total checkpoint contents: **3,160 files**, **927.29 GiB** (995,673,068,120 bytes).
+
+```text
+continuous/
+  qwen3base/<vision-encoder>/<training-stage>/...
+  smollm2/<vision-encoder>/<training-stage>/...
+  qwen3/<vision-encoder>/<training-stage>/...
+  qwen25/<vision-encoder>/<training-stage>/...
+FILES.json
+```
+
+Only the four completely restored continuous groups are included. Incomplete discrete archive volumes are excluded.
+
+## Download
+
+```python
+from huggingface_hub import snapshot_download
+
+snapshot_download(
+    repo_id="336labs/VisionEncoder-MLLM-Eval",
+    allow_patterns=["continuous/qwen3base/**", "FILES.json", "README.md"],
+    local_dir="VisionEncoder-MLLM-Eval",
+)
+```
+
+Choose the needed language-backbone group and vision-encoder run. Each run retains its original weights, configuration, tokenizer files, and available training metadata. Check the configuration in that subdirectory for the corresponding architecture.