llama.cpp

10,108 Commits 4 Branches 0 Tags

2 Commits

Author	SHA1	Message	Date
Yash Raj PandeyandGitHub	4937ca83f4	llama-quant : exclude i32 ffn_gate_tid2eid routing table from quantization (#25787 ) DeepSeek-V4's ffn_gate_tid2eid tensor is an i32 token-id -> expert-id index table, not weights. It was never added to the name-based exclusion list alongside ffn_gate_inp.weight, so llama-quantize tries to quantize it and fails since i32 cannot convert to a float type. Fixes ggml-org/llama.cpp#25754 Signed-off-by: Yash Raj Pandey <yashpn62@gmail.com>	2026-07-18 13:43:18 +02:00
Yash Raj Pandey GitHub Georgi Gerganov	fd3271e0b4	ggml-cpu : fix rms_norm_back wrong output under in-place aliasing (#24305 ) * ggml-cpu : fix rms_norm_back wrong output under in-place aliasing * cont : clean-up comment --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>	2026-06-09 10:24:27 +03:00

Author

SHA1

Message

Date

Yash Raj PandeyandGitHub

4937ca83f4

llama-quant : exclude i32 ffn_gate_tid2eid routing table from quantization (#25787 )

DeepSeek-V4's ffn_gate_tid2eid tensor is an i32 token-id -> expert-id
index table, not weights. It was never added to the name-based
exclusion list alongside ffn_gate_inp.weight, so llama-quantize tries
to quantize it and fails since i32 cannot convert to a float type.

Fixes ggml-org/llama.cpp#25754

Signed-off-by: Yash Raj Pandey <yashpn62@gmail.com>

2026-07-18 13:43:18 +02:00

fd3271e0b4

ggml-cpu : fix rms_norm_back wrong output under in-place aliasing (#24305 )

* ggml-cpu : fix rms_norm_back wrong output under in-place aliasing

* cont : clean-up comment

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

2026-06-09 10:24:27 +03:00