Upload README.md with huggingface_hub

README.md

@@ -20,39 +20,28 @@ Ivan Lee, Cheng Yang, Taylor Berg-Kirkpatrick
 
 ## Available Checkpoints
 
-### Vision (base, 768x768)
+Naming convention: `{regime}_{config}_h{N}_{objective}[_recon-init]`
 
-| Checkpoint | Objective | Hybrid | Training | CR | PPL |
-|------------|-----------|--------|----------|-----|-----|
-| `vision_base_h0_recon` | Reconstruction | 0 | - | 3.60 | 1.03 |
-| `vision_base_h0_lm` | LM | 0 | Direct | 3.60 | 5.08 |
-| `vision_base_h0_lm_recon-init` | LM | 0 | From recon | 3.60 | 5.06 |
+### Reconstruction
 
-### Meanpool (w4s4)
+| Checkpoint | Regime | CR | PPL |
+|------------|--------|-----|-----|
+| `vision_base_h0_recon` | Vision base | 3.60 | 1.03 |
+| `meanpool_w4s4_h0_recon` | Meanpool w4s4 | 3.97 | 1.04 |
 
-| Checkpoint | Objective | Hybrid | Training | CR | PPL |
-|------------|-----------|--------|----------|-----|-----|
-| `meanpool_w4s4_h0_recon` | Reconstruction | 0 | - | 3.97 | 1.04 |
-| `meanpool_w4s4_h0_lm_recon-init` | LM | 0 | From recon | 3.97 | 5.02 |
+### Language Modeling
 
-## Naming Convention
-
-```
-{regime}_{config}_h{N}_{objective}[_recon-init]
-```
-
-| Field | Values | Description |
-|-------|--------|-------------|
-| regime | vision, conv1d, meanpool, text | Compression architecture |
-| config | base/small/tiny/large, t500/t250, w4s4/w10s10, ctx525 | Regime-specific config |
-| h{N} | h0, h100 | Hybrid text tokens (0 = pure vision/compression) |
-| objective | recon, lm | Training objective |
-| recon-init | (optional) | LM initialized from reconstruction checkpoint |
+| Checkpoint | Regime | CR | Init | PPL |
+|------------|--------|-----|------|-----|
+| `vision_base_h0_lm` | Vision base | 3.60 | Direct | 5.08 |
+| `vision_base_h0_lm_recon-init` | Vision base | 3.60 | From recon | 5.06 |
+| `meanpool_w4s4_h0_lm_recon-init` | Meanpool w4s4 | 3.97 | From recon | 5.02 |
 
 ## Model Details
 
-- **Architecture**: DeepSeek-OCR with trainable vision encoder
-- **Encoder Status**: Trained (not frozen)
+- **Architecture**: DeepSeek-OCR with vision encoder
+- **Vision checkpoints**: Trained encoder, 768x768 (base)
+- **Meanpool checkpoints**: Frozen encoder, window=4, stride=4
 - **Dataset**: 510k samples from FineWiki
 
 ## Usage