Link required vLLM PRs (#52184, #55133) in model card

README.md

@@ -71,11 +71,16 @@ vllm serve thinkingmachines/Inkling-Small-NVFP4 \
     --trust-remote-code
 ```
 
-> **Note:** Inkling currently requires `--enforce-eager` (CUDA-graph capture of
-> the custom ops crashes). The drafter uses the full padded vocab (draft
-> `vocab_size` 201024 vs. the verifier's logical 200058); loading it in vLLM
-> needs [vllm-project/vllm#55133](https://github.com/vllm-project/vllm/pull/55133),
-> which treats a padded/full-vocab draft as an identity map (no `d2t` required).
+> **Required vLLM patches** (both open at time of writing):
+> - [vllm-project/vllm#52184](https://github.com/vllm-project/vllm/pull/52184) —
+>   adds Eagle3 hidden-state extraction support for Inkling (the verifier must
+>   expose the aux hidden states this drafter consumes).
+> - [vllm-project/vllm#55133](https://github.com/vllm-project/vllm/pull/55133) —
+>   treats a full/padded-vocab draft (draft `vocab_size` 201024 vs. the verifier's
+>   logical 200058) as an identity map, so no `d2t` mapping is required to load it.
+>
+> Also note: Inkling currently requires `--enforce-eager` (CUDA-graph capture of
+> the custom ops crashes).
 
 ## Training