vulkan upload weights per layer, reduce peak ram usage

[nihui]

No description provided.

[tencent-adm]

Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[Copilot]

Pull request overview

This pull request refactors Vulkan model weight uploading to reduce peak RAM usage by uploading weights layer-by-layer instead of all at once. The key optimization is chunking uploads into 256MB batches, which reduces memory pressure during model loading.

Changes:

• Removed centralized NetPrivate::upload_model() method and moved upload logic inline into Net::load_model() to enable per-layer uploads
• Added VkTransfer::reset() and VkTransfer::pending_upload_total() methods to support incremental uploads with periodic submission
• Implemented upload chunking logic that submits and resets the VkTransfer command buffer after accumulating 256MB of pending uploads

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
src/net.cpp	Moved weight upload logic inline with model loading loop; added 256MB chunking to reduce peak memory usage
src/command.h	Added reset() and pending_upload_total() methods to VkTransfer class API
src/command.cpp	Implemented upload tracking, reset functionality, and getter for pending upload size

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 45.61404% with 31 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.94%. Comparing base (e665c48) to head (53a3d99).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
src/command.cpp	18.51%	22 Missing ⚠️
src/net.cpp	70.00%	9 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6534      +/-   ##
==========================================
- Coverage   92.96%   92.94%   -0.02%     
==========================================
  Files         809      809              
  Lines      256977   256643     -334     
==========================================
- Hits       238900   238546     -354     
- Misses      18077    18097      +20     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.