f8d4abae86
convert : support Qwen3.5/Qwen3.5 Moe NVFP4 and add input scales ( #20505 )
...
* convert : fix Qwen3.5 NVFP4 conversion
* Updated copilot concerns and rebased
* move into _LinearAttentionVReorderBase and simplify
* --flake
* new_name not needed
* Added input_scale to gguf
* Fixed input_scale addition as tensor
* Added input scale to loader and named _in_s
* Update convert_hf_to_gguf.py
Re-removed input_scale from aux cleanup
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com >
---------
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com >
2026-03-26 16:52:06 +01:00
112c78159f
ggml-cuda: Add NVFP4 dp4a kernel ( #20644 )
...
Added check for dst_t to cuda_cast template for float
Restored ggml_cuda_ue4m3_to_fp32, changed vecdot ints to int32ts
Added CUDART/HIP Check and HIP/fp8 include
Added NVFP4 to Test-backend-ops
Added hip_fp8_e4m3 to __nv_fp8_e4m3 typedef
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de >
2026-03-26 09:54:03 +01:00
eac9c6ea83
Convert: Make NVFP4 and MXFP4 HF conversions say NVFP4/MXFP4 instead of BF16 ( #20730 )
...
* Corrected convert script for NVFP4 naming and updated gguf constants
* Add mostly_MXFP4 to FileType
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com >
* Update convert_hf_to_gguf.py
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com >
* Update convert_hf_to_gguf.py
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com >
* simplify
* set initial value [no ci]
---------
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com >
2026-03-21 13:35:21 +02:00
Michael Wand and GitHub
d23355afc3
model : wire up Qwen3.5/Qwen3.5MoE tensors for NVFP4 support ( #20506 )
2026-03-14 22:44:42 +01:00
Michael Wand and GitHub
600a366478
Corrected: changed s13 = src1->nb[3] instead of nb[2] ( #18724 )
2026-01-10 10:16:07 +01:00
