Update README.md

README.md

@@ -12,4 +12,48 @@ tags:
 base_model:
 - nvidia/Cosmos-Predict2-14B-Text2Image
 base_model_relation: quantized
----
+---
+For more information (including how to compress models yourself), check out https://huggingface.co/DFloat11 and https://github.com/LeanModels/DFloat11
+
+Feel free to request for other models for compression as well, although models whose architecture I am unfamiliar with might be slightly tricky for me.
+
+### How to Use
+
+#### ComfyUI
+Install my own fork of the DF11 ComfyUI custom node: https://github.com/mingyi456/ComfyUI-DFloat11-Extended instead. After installing the DF11 custom node, use the provided workflow [json](cosmos_predict2_14B_t2i-DF11-workflow.json), or simply replace the "Load Diffusion Model" node of an existing Kontext workflow with the "DFloat11 Model Loader" node. If you run into any issues, feel free to leave a comment. The workflow is also embedded in the below [png](cosmos_predict2_14B_t2i-DF11-workflow.png) image.
+
+![](cosmos_predict2_14B_t2i-DF11-workflow.png)
+
+#### `diffusers`
+Refer to this [model](https://huggingface.co/mingyi456/Cosmos-Predict2-14B-Text2Image-DF11) instead.
+
+### Compression Details
+
+This is the `pattern_dict` for compression:
+
+```python
+pattern_dict_comfyui = {
+        "t_embedder\.1": (
+            "linear_1",
+            "linear_2",
+        ),
+        r"blocks\.\d+": (
+            "self_attn.q_proj",
+            "self_attn.k_proj",
+            "self_attn.v_proj",
+            "self_attn.output_proj",
+            "cross_attn.q_proj",
+            "cross_attn.k_proj",
+            "cross_attn.v_proj",
+            "cross_attn.output_proj",
+            "mlp.layer1",
+            "mlp.layer2",
+            "adaln_modulation_self_attn.1",
+            "adaln_modulation_self_attn.2",
+            "adaln_modulation_cross_attn.1",
+            "adaln_modulation_cross_attn.2",
+            "adaln_modulation_mlp.1",
+            "adaln_modulation_mlp.2",
+        )
+    }
+```