Skip to content

Added NPI files#837

Open
asmigosw wants to merge 3 commits intoquic:mainfrom
asmigosw:NPI_configs_to_mainline
Open

Added NPI files#837
asmigosw wants to merge 3 commits intoquic:mainfrom
asmigosw:NPI_configs_to_mainline

Conversation

@asmigosw
Copy link
Contributor

@asmigosw asmigosw commented Mar 9, 2026

Added NPI files for models:

  • Qwen/Qwen2.5-VL-32B-Instruct-AWQ
  • openai/whisper-large-v3-turbo
  • facebook/wav2vec2-base-960h
  • Qwen/Qwen3-30B-A3B-Instruct-2507

Signed-off-by: Asmita Goswami <asmigosw@qti.qualcomm.com>
@@ -0,0 +1,2 @@
FP16NodeInstanceNames: ['onnx::MatMul_1328','/wav2vec2/feature_projection/projection/MatMul_output_0', 'onnx::MatMul_1401', '/wav2vec2/encoder/layers.5/attention/q_proj/MatMul_output_0', 'onnx::MatMul_1408', '/wav2vec2/encoder/layers.5/attention/k_proj/MatMul_output_0', 'onnx::MatMul_1409', '/wav2vec2/encoder/layers.5/attention/v_proj/MatMul_output_0', 'onnx::MatMul_1411', '/wav2vec2/encoder/layers.5/attention/out_proj/MatMul_output_0', 'onnx::MatMul_1412', '/wav2vec2/encoder/layers.5/feed_forward/intermediate_dense/MatMul_output_0', 'onnx::MatMul_1413', '/wav2vec2/encoder/layers.5/feed_forward/output_dense/MatMul_output_0', 'onnx::MatMul_1414', '/wav2vec2/encoder/layers.6/attention/q_proj/MatMul_output_0', 'onnx::MatMul_1421', '/wav2vec2/encoder/layers.6/attention/k_proj/MatMul_output_0', 'onnx::MatMul_1422', '/wav2vec2/encoder/layers.6/attention/v_proj/MatMul_output_0', 'onnx::MatMul_1424', '/wav2vec2/encoder/layers.6/attention/out_proj/MatMul_output_0', 'onnx::MatMul_1425', '/wav2vec2/encoder/layers.6/feed_forward/intermediate_dense/MatMul_output_0', 'onnx::MatMul_1426', '/wav2vec2/encoder/layers.6/feed_forward/output_dense/MatMul_output_0']
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Lets add these NPI files in the corresponding modelling folder too, similar to Gemma3.
Also you may have to edit the config file which is pointing to default NPI to be picked. refer PR#810

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants