Skip to content

Qwen3Vl#801

Open
qcdipankar wants to merge 17 commits intomainfrom
qwen3_vl_mainline
Open

Qwen3Vl#801
qcdipankar wants to merge 17 commits intomainfrom
qwen3_vl_mainline

Conversation

@qcdipankar
Copy link
Contributor

No description provided.

@qcdipankar qcdipankar force-pushed the qwen3_vl_mainline branch 4 times, most recently from 5bc0eb7 to 19a163b Compare March 2, 2026 06:37
#
# -----------------------------------------------------------------------------

from time import perf_counter
Copy link
Contributor

Choose a reason for hiding this comment

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

Follow the same structure while creating the examples https://github.com/quic/efficient-transformers/blob/main/examples/CONTRIBUTING.md

Copy link
Contributor Author

Choose a reason for hiding this comment

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

in case of qwen3 we need to have different example files

height=354,
width=536,
mxfp6_matmul=True,
mxint8_kv_cache=True,
Copy link
Contributor

Choose a reason for hiding this comment

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

subfunction

if output_name.endswith("_RetainedState"):
custom_io_lang[output_name[: -len("_RetainedState")]] = (
"float16" if "vision_embeds" in output_name else kv_cache_dtype
"float16"
Copy link
Contributor

Choose a reason for hiding this comment

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

can you take it from a top level argument passed from pretrained()

qcdipankar and others added 13 commits March 11, 2026 03:24
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <quic_dipankar@quicinc.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Onkar Chougule <ochougul@qti.qualcomm.com>
Co-authored-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <quic_dipankar@quicinc.com>
Signed-off-by: Onkar Chougule <ochougul@qti.qualcomm.com>
Signed-off-by: Dhiraj Kumar Sah <dhirajku@qti.qualcomm.com>
Co-authored-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Co-authored-by: Onkar Chougule <168134249+ochougul@users.noreply.github.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
The Onboarding of Qwen3VlMoe

---------

Signed-off-by: Dipankar Sarkar <quic_dipankar@quicinc.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: vtirumal <vtirumal@qti.qualcomm.com>
Signed-off-by: Onkar Chougule <ochougul@qti.qualcomm.com>
Co-authored-by: vtirumal <vtirumal@qti.qualcomm.com>
Co-authored-by: Onkar Chougule <168134249+ochougul@users.noreply.github.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <quic_dipankar@quicinc.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: vtirumal <vtirumal@qti.qualcomm.com>
Co-authored-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Added Support for Subfn for Qwen 3 VL dense, MOE.
Updated prefill only logic for disagg mode

---------

Signed-off-by: vtirumal <vtirumal@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>
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.

6 participants