Tags

Tags give the ability to mark specific points in history as being important

b2144

ea9c8e11 · llama : add support for Nomic Embed (#5468) · Feb 13, 2024
b2143

c4e6dd59 · llama : allow raw byte in SPM vocabs; don't crash on nl 404 (#5478) · Feb 13, 2024
b2142

037259be · llama : make load error reporting more granular (#5477) · Feb 13, 2024
b2141

26397890 · finetune : rename feed-forward tensors (w1/w2/w3) (#4839) · Feb 13, 2024
b2140

cf45252a · tests : multi-thread the tokenizer tests (#5474) · Feb 13, 2024
b2139

03bf161e · llama : support batched embeddings (#5466) · Feb 13, 2024
b2138

ad014bba · make: add error message for bad CUDA version (#5444) · Feb 13, 2024
b2137

49cc1f7d · bert : add tests + fix quantization (#5475) · Feb 13, 2024
b2136

99b8b43d · tests : disable moe test (#5473) · Feb 13, 2024
b2135

895407f3 · ggml-quants : fix compiler warnings (shadow variable) (#5472) · Feb 13, 2024
b2134

099afc62 · llama : fix quantization when tensors are missing (#5423) · Feb 12, 2024
b2133

df334a11 · swift : package no longer use ggml dependency (#5465) · Feb 12, 2024
b2131

43fe07c1 · ggml-sycl: Replace 3d ops with macro (#5458) · Feb 12, 2024
b2130

4a46d2b7 · llava : remove prog parameter from ArgumentParser (#5457) · Feb 12, 2024
b2129

3b169441 · sync : ggml (#5452) · Feb 12, 2024
b2128

3bdc4cd0 · CUDA: mul_mat_vec_q tiling, refactor mul mat logic (#5434) · Feb 11, 2024
b2127

2891c8aa · Add support for BERT embedding models (#5423) · Feb 11, 2024
b2125

c88c74f9 · vulkan: only use M-sized matmul on Apple GPUs (#5412) · Feb 11, 2024
b2124

a803333a · common : use enums for sampler types (#5418) · Feb 11, 2024
b2123

68478014 · server : allow to specify tokens as strings in logit_bias (#5003) · Feb 11, 2024

Prev
1
…
25
26
27
28
29
30
31
32
33
…
99
Next