Tags

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

b2397

ecab1c75 · cmake : fix subdir for `LLAMA_METAL_EMBED_LIBRARY` (#5985) · Mar 11, 2024
b2396

ee35600b · llama : fix F16/F32 downcast + improve names (#5980) · Mar 11, 2024
b2395

be858f62 · Better 1.5 bit quantization (#5971) · Mar 11, 2024
b2394

ef3ced26 · [SYCL] Add q3_s and q1_s (#5886) · Mar 11, 2024
b2393

3814a073 · [SYCL] Add support for SYCL Nvidia target (#5738) · Mar 11, 2024
b2392

bb6d00bb · metal : move mm_id indices to shared mem (#5982) · Mar 10, 2024
b2391

7ab7b733 · android : fix utf8 decoding error (#5935) · Mar 10, 2024
b2389

b838b53a · sync : ggml · Mar 10, 2024
b2387

bf47a5ee · ggml : remove __constant__ specifier for CUDA tables (#5940) · Mar 10, 2024
b2386

fa8a809a · server: ci: windows build and tests (#5968) · Mar 10, 2024
b2385

bcebd7db · llama : add support for GritLM (#5959) · Mar 10, 2024
b2384

2960eae8 · grammar : verify parsed state (#5950) · Mar 10, 2024
b2382

621e86b3 · server: benchmark: chat/completions scenario and other llm servers comparison (#5941) · Mar 09, 2024
b2381

77d1ac7e · server : print chat template info · Mar 09, 2024
b2380

d894f352 · perplexity : support using multiple sequences to allow larger batch sizes (#5946) · Mar 09, 2024
b2378

8380ecfb · ggml : fix unnecessary f32 -> f16 -> f32 casts (mmla) (#5951) · Mar 09, 2024
b2377

58308a0e · server : fix metrics init (#5964) · Mar 09, 2024
b2376

5b097973 · ggml : remove old quantization functions (#5942) · Mar 09, 2024
b2374

fb215c38 · server : normalize embeddings (#5956) · Mar 09, 2024
b2372

0db32bea · server : fix passing prompt as tokens (#5955) · Mar 09, 2024

Prev
1
…
17
18
19
20
21
22
23
24
25
…
99
Next