Add CRISPR 7B/16x checkpoint; update model card

7b_16x/checkpoint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aebf530a60a1e42d7bc58c35c84eee7fb6439a98efec7993aaedfd5db73d3482
+size 1541964962

7b_16x/config.json

@@ -0,0 +1,40 @@
+{
+  "decoder_path": "./Qwen/Qwen2.5-VL-7B-Instruct",
+  "use_token_mixer": true,
+  "token_mixer_num_layers": 2,
+  "token_mixer_num_heads": 16,
+  "token_mixer_dropout": 0.1,
+  "freeze_token_mixer": false,
+  "use_local_c3": true,
+  "local_c3_block_size": 4,
+  "local_c3_num_heads": 8,
+  "num_post_encoder_layers": 2,
+  "use_rope_alignment": true,
+  "use_kl_distillation": true,
+  "kl_loss_weight": 0.5,
+  "teacher_temperature": 1.0,
+  "use_hidden_distillation": true,
+  "hidden_loss_weight": 0.1,
+  "use_decoder_lora": false,
+  "data_path": "./data/train/stage2_mixed/combined_v4.jsonl",
+  "max_length": 768,
+  "max_samples": null,
+  "max_image_size": 672,
+  "output_dir": "./outputs/image_c3_v7_16x_stage2_20260311_150346",
+  "epochs": 3,
+  "batch_size": 1,
+  "gradient_accumulation_steps": 16,
+  "lr_local_c3": 1e-05,
+  "lr_token_mixer": 1e-05,
+  "lr_decoder_lora": 1e-05,
+  "weight_decay": 0.01,
+  "warmup_ratio": 0.1,
+  "max_grad_norm": 1.0,
+  "log_steps": 50,
+  "save_steps": 500,
+  "gpu": 0,
+  "num_workers": 4,
+  "resume_from": null,
+  "timestamp": "2026-03-11T19:46:38.538636",
+  "world_size": 4
+}

README.md

@@ -1,13 +1,15 @@
 ---
 license: mit
-base_model: Qwen/Qwen2.5-VL-3B-Instruct
+base_model:
+- Qwen/Qwen2.5-VL-3B-Instruct
+- Qwen/Qwen2.5-VL-7B-Instruct
 tags:
 - vision-language-model
 - token-compression
 - multimodal
 ---
 
-# CRISPR — Checkpoints (Qwen2.5-VL-3B backbone)
+# CRISPR — Checkpoints
 
 Checkpoints for **CRISPR: Context-Refined Information Spatial Pooling with
 Region-awareness for Efficient Visual Token Compression in VLMs**, accepted
@@ -16,9 +18,10 @@ at ACM MM 2026.
 - Code: https://github.com/ZuyiZhou/CRISPR
 - Paper DOI: https://doi.org/10.1145/3767308.3835007
 
-This repo currently hosts the **Qwen2.5-VL-3B-Instruct** backbone checkpoints
-at 9x and 16x compression. Qwen2.5-VL-7B backbone checkpoints will be added
-in a subsequent update.
+This repo hosts CRISPR checkpoints for the Qwen2.5-VL-3B-Instruct backbone
+(9x and 16x compression) and the Qwen2.5-VL-7B-Instruct backbone (16x
+compression). The 7B/9x checkpoint is not currently available (lost prior to
+this release) and is not planned unless retraining happens in the future.
 
 ## Files
 
@@ -28,6 +31,8 @@ in a subsequent update.
 | `3b_9x/config.json` | | | training config used to produce this checkpoint |
 | `3b_16x/checkpoint.pt` | 16x (4x4 block) | Qwen2.5-VL-3B-Instruct | Stage-2, best checkpoint by val loss |
 | `3b_16x/config.json` | | | training config used to produce this checkpoint |
+| `7b_16x/checkpoint.pt` | 16x (4x4 block) | Qwen2.5-VL-7B-Instruct | Stage-2, best checkpoint by val loss |
+| `7b_16x/config.json` | | | training config used to produce this checkpoint |
 
 Each `checkpoint.pt` is a plain `torch.save` dict with keys `config`,
 `token_mixer` (TokenMixer state dict), and `local_c3` (LocalC3 state dict,
@@ -45,6 +50,7 @@ from crispr import create_model_v7
 
 model = create_model_v7(decoder_path="./Qwen/Qwen2.5-VL-3B-Instruct")
 model.load_checkpoint("3b_9x/checkpoint.pt")  # see crispr/model_v7.py for the loader
+# for the 7B backbone: decoder_path="./Qwen/Qwen2.5-VL-7B-Instruct", checkpoint="7b_16x/checkpoint.pt"
 ```
 
 See the main repository (https://github.com/ZuyiZhou/CRISPR) for the model