llama.cpp
(MIRROR) Implement gguf init from buffer
#104
Open

(MIRROR) Implement gguf init from buffer #104

ngxson wants to merge 14 commits into ngxson:master from giladgd:implementGgufInitFromBuffer
ngxson
giladgd ggml: implement `gguf_init_from_buffer`
9b2aa8c9
giladgd test: `gguf_init_from_buffer`
de9ebec6
giladgd fix: memory breakdown for a model loaded with `no_alloc` from a file …
59883b3c
giladgd fix: use `GGML_UNUSED`
5950daa7
giladgd fix: remove `total_size` from `gguf_reader`
b120c9bd
giladgd fix: file offset calculation, rename `offset` to `data_offset`
65f71360
giladgd refactor: extract model loader bug fixes to another PR
0a59d4e2
giladgd feat: add `gguf_init_from_callback`
62f4e953
giladgd fix: always require a max expected size
78bff38f
giladgd fix: change `gguf_reader_callback_t`'s `output` type to `void *`, cha…
dcca71d3
giladgd fix: harden against offset overflow in buffer read
6248d3cd
coderabbitai
github-actions github-actions added ggml
github-actions github-actions added testing
coderabbitai
coderabbitai commented on 2026-05-11
giladgd fix: remove seek behavior from the callback
913a28f5
coderabbitai
coderabbitai commented on 2026-05-21
giladgd feat: `max_chunk_read == 0` means `SIZE_MAX`
7157cc0a
giladgd fix: seeking in a gguf file with no tensors
5908c589

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone