Fix Llama importer (NVIDIA-NeMo#13583)

suiyoubi · web-flow · commit b912a85cf560 · 2025-05-15T02:31:40.000Z
* Fix Llama importer

Signed-off-by: Ao Tang &lt;aot@nvidia.com&gt;

* Fix Llama importer

Signed-off-by: Ao Tang &lt;aot@nvidia.com&gt;

---------

Signed-off-by: Ao Tang &lt;aot@nvidia.com&gt;
diff --git a/nemo/collections/llm/gpt/model/llama.py b/nemo/collections/llm/gpt/model/llama.py
@@ -765,7 +765,7 @@ def make_vocab_size_divisible_by(vocab_size):
             params_dtype=dtype_from_hf(source),
             generation_config=generation_config,
             vocab_size=source.vocab_size,
-            kv_channels=getattr(source, "head_dim"),
+            kv_channels=getattr(source, "head_dim", None),
             **args,
         )
 

Original file line number	Diff line number	Diff line change
`@@ -765,7 +765,7 @@ def make_vocab_size_divisible_by(vocab_size):`
`765`	`765`	`params_dtype=dtype_from_hf(source),`
`766`	`766`	`generation_config=generation_config,`
`767`	`767`	`vocab_size=source.vocab_size,`
`768`		`- kv_channels=getattr(source, "head_dim"),`
	`768`	`+ kv_channels=getattr(source, "head_dim", None),`
`769`	`769`	`**args,`
`770`	`770`	`)`
`771`	`771`