[Model] Add Afmoe architecture implementation (vllm-project#28332)

Signed-off-by: Maziyar Panahi <[email protected]>
Signed-off-by: Pranav <[email protected]>
Co-authored-by: Maziyar Panahi <[email protected]>

Author: pranav4501

docs/models/supported_models.md

@@ -351,6 +351,7 @@ th {
 
 | Architecture | Models | Example HF Models | [LoRA](../features/lora.md) | [PP](../serving/parallelism_scaling.md) |
 |--------------|--------|-------------------|----------------------|---------------------------|
+| `AfmoeForCausalLM` | Afmoe | TBA | ✅︎ | ✅︎ |
 | `ApertusForCausalLM` | Apertus | `swiss-ai/Apertus-8B-2509`, `swiss-ai/Apertus-70B-Instruct-2509`, etc. | ✅︎ | ✅︎ |
 | `AquilaForCausalLM` | Aquila, Aquila2 | `BAAI/Aquila-7B`, `BAAI/AquilaChat-7B`, etc. | ✅︎ | ✅︎ |
 | `ArceeForCausalLM` | Arcee (AFM) | `arcee-ai/AFM-4.5B-Base`, etc. | ✅︎ | ✅︎ |

tests/models/registry.py

@@ -173,6 +173,10 @@ def check_available_online(
 
 _TEXT_GENERATION_EXAMPLE_MODELS = {
     # [Decoder-only]
+    "AfmoeForCausalLM": _HfExamplesInfo(
+        "arcee-ai/Trinity-Nano",
+        is_available_online=False,
+    ),
     "ApertusForCausalLM": _HfExamplesInfo("swiss-ai/Apertus-8B-Instruct-2509"),
     "AquilaModel": _HfExamplesInfo("BAAI/AquilaChat-7B", trust_remote_code=True),
     "AquilaForCausalLM": _HfExamplesInfo("BAAI/AquilaChat2-7B", trust_remote_code=True),