Skip to content
GitLab
Explore
Sign in
Tags
Tags give the ability to mark specific points in history as being important
b1971
1182cf4d
·
Another bucket sort (#5109)
·
Jan 26, 2024
b1969
5eaf9964
·
llama : dynamic temperature sampling (#4972)
·
Jan 25, 2024
b1966
faa3526a
·
Fix Q3_K_XS for MoE models (#5113)
·
Jan 25, 2024
b1965
ddc5a503
·
metal : show compile log messages
·
Jan 25, 2024
b1964
cd4fddb2
·
cuda : fix 2-bit quants on amd hip (#5105)
·
Jan 24, 2024
b1961
1387ea21
·
llama : pre-allocate input tensors in a separate buffer (#5100)
·
Jan 24, 2024
b1960
26d60760
·
metal : disable support for MUL_MAT F32 x F16
·
Jan 23, 2024
b1959
44879ee8
·
Additional KL-divergence statistics (#5081)
·
Jan 23, 2024
b1958
9ecdd12e
·
CUDA: more info when no device code (#5088)
·
Jan 23, 2024
b1957
89758723
·
minor : clean-up some warnings and style (#5094)
·
Jan 23, 2024
b1956
2bed4aa3
·
devops : add intel oneapi dockerfile (#5068)
·
Jan 23, 2024
b1954
011e8ec5
·
llama : fix not enough space in buffer with Qwen (#5086)
·
Jan 22, 2024
b1953
6f9939d1
·
KL-divergence (#5076)
·
Jan 22, 2024
b1952
780e24a2
·
ggml : parallelize FP32 conversion when using BLAS (#5045)
·
Jan 22, 2024
b1951
3ce7e8f8
·
llava : MobileVLM support (#4954)
·
Jan 22, 2024
b1950
b2d80e10
·
flake.nix: add a comment about flakes vs nix
·
Jan 22, 2024
b1943
15bceec2
·
imatrix : keep intermediate imatrix results (#5077)
·
Jan 22, 2024
b1942
d6bd4d46
·
llama : support StableLM 2 1.6B (#5052)
·
Jan 22, 2024
b1941
152d9d05
·
finetune : print sample-start/include-sample-start (#5072)
·
Jan 22, 2024
b1940
66d575c4
·
llama : add Q3_K_XS (#5060)
·
Jan 22, 2024
Prev
1
…
31
32
33
34
35
36
37
38
39
…
99
Next