finetune: Add -ngl parameter #3762
ggerganov
merged
16 commits into
ggml-org:master
from
AndrewGodfrey:finetune_enableGpu
Add '-ngl' support to finetune.cpp
facb1a3e
Add fprintf in ggml_cuda_op_add
4d452dbc
Add 'finetune.sh', which currently fails when using GPU
e1ebce03
tweak finetune.sh
1758d0ab
Suppress some warnings in ggml.c
9ea91cea
Add f16 implementation to ggml_compute_forward_add_f16_f32
19097c97
Add an f16 case to ggml_add_cast_impl and llama_build_lora_finetune_g…
7cbf5b28
finetune.sh: Edit comments
9587ab4c
Add "add_f16_f32_f32_cuda"
86ceda42
Tweak an error message
81dabd8e
cebtenzzre
commented
on 2023-10-24
finetune.sh: Add an optional LLAMA_MODEL_DIR variable
6359c151
finetune.sh: Add an optional LLAMA_TRAINING_DIR variable
fe44ded0
kotee4ko
approved these changes
on 2023-10-27
train : minor
c5c54d10
ggerganov
approved these changes
on 2023-10-28
tabs to spaces
998a548a
Merge branch 'master' of https://github.com/AndrewGodfrey/llama.cpp i…
653bc1c0
Merge branch 'master' of https://github.com/ggerganov/llama.cpp into …
3af77560
ggerganov
approved these changes
on 2023-11-01
Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub