Dummy-Test-PR

[qraniumcitest]

No description provided.

[qraniumcitest]

Jenkins-Ready

[qraniumcitest]

Build started: https://re-qranium-ci.qualcomm.com/job/Qeff-Q1/job/Qeff-Pipeline/39/

[qraniumcitest]

From GHES (comment) by @

Code Assistant

Reviewed Commits: 9ec964b, cb7da87, 18e39a3, 0c1134d, ce71a04, 06d8f88, 96b7379, 6b2a01b, eb45f6f, babc96d, 009a778

• 9ec964b: Create Mirror_Fork_PR_To_Ghes.yml

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• cb7da87: updated notebooks (updated notebooks quic/efficient-transformers#543)

Updated the correct code with updated syntax, removed device_group
parameter in model.compile()

Signed-off-by: Sharvari Medhe smedhe@qti.qualcomm.com

• 18e39a3: Update Mirror_Fork_PR_To_Ghes.yml

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• 0c1134d: Update execute.py

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• ce71a04: Update qnn_compiler.py

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• 06d8f88: Update compile.py

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• 96b7379: Update qnn_config.json

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• 6b2a01b: Create Mirror_Fork_PRs_to_GHES.yml

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• eb45f6f: Update execute.py

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

• babc96d: Merge branch 'main' into jenkins-Ready-Test
• 009a778: Update compile.py

Signed-off-by: qraniumcitest rmakar@qti.qualcomm.com

This PR introduces several changes across multiple files, primarily focusing on:

1. Adding a GitHub workflow for mirroring fork PRs to GHES (GitHub Enterprise Server)
2. Code cleanup in QEfficient modules
3. Updates to notebook examples for tokenizer integration
4. Parameter name standardization (mxfp6 → mxfp6_matmul)

File	Changes	Issues	Highest Severity
.github/workflows/Mirror_Fork_PR_To_GHES.yml	New file	1	Medium
.github/workflows/Mirror_Fork_PRs_to_GHES.yml	New file	1	Medium
QEfficient/cloud/compile.py	Minor	0	-
QEfficient/cloud/execute.py	Minor	0	-
QEfficient/compile/qnn_compiler.py	Minor	0	-
QEfficient/compile/qnn_config.json	Minor	0	-
notebooks/QEfficientGPT2.ipynb	Updates	1	Medium
notebooks/QEfficientMPT.ipynb	Updates	0	-

The most significant issues are:

1. Duplicate GitHub workflow files with nearly identical content
2. Parameter name inconsistency in notebook examples vs implementation
3. Minor whitespace and newline issues

[DUPLICATION] Duplicate GitHub workflow files with nearly identical content - Medium

There are two nearly identical GitHub workflow files: Mirror_Fork_PR_To_GHES.yml and Mirror_Fork_PRs_to_GHES.yml. These files contain almost the same content with only minor differences (e.g., one has a whitespace trim function for user names). Having duplicate workflow files can lead to confusion, maintenance issues, and unexpected behavior.

Recommendation: Keep only one of these files (preferably the more complete one with the whitespace trim functionality) and remove the other.

[CONSISTENCY] Parameter name inconsistency between notebooks and implementation - Medium

The notebooks (QEfficientGPT2.ipynb and QEfficientMPT.ipynb) use mxfp6_matmul=True as a parameter when calling compile(), but the implementation in qnn_compiler.py uses mxfp6=True. This inconsistency can lead to confusion and potential errors.

The parameter name should be consistent across all files. Based on the implementation in qnn_compiler.py (line 332), the parameter should be mxfp6, not mxfp6_matmul.

Fixed Code Snippet:

qeff_model.compile(num_cores=14, mxfp6=True)

[TYPO] Typo in error message in GitHub workflow - Low

There's a typo in the error message in both GitHub workflow files. The message says "Connect VPN then re-reun" instead of "Connect VPN then re-run".

Fixed Code Snippet:

curl -I --max-time 10 "https://github.qualcomm.com" ||  (echo "Cannot reach GHES over HTTPS. Connect VPN then re-run"; exit 1)

---

** Version 1.3.2**

Help us improve!

How useful was this code feedback? Not very useful 1️⃣ 2️⃣ 3️⃣ 4️⃣ 5️⃣ Very useful

How much time did it save you (in hours)? 0 | <1 | 1-3 | >4

Let us know your detailed feedback

Source: https://github.qualcomm.com/qranium/efficient-transformers/pull/31#issuecomment-988408

[qraniumcitest]

Jenkins-Ready

[qraniumcitest]

Build started: https://re-qranium-ci.qualcomm.com/job/Qeff-Q1/job/Qeff-Pipeline/40/