HappyZ happyz
0 Followers · 0 Following
Repositories
18
 Projects Packages Public Activity Starred Repositories
happyz synced commits to refs/pull/6773/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:01 -07:00
9f315b6865 Merge ac6ae5daca into c70bfd7bcb
c70bfd7bcb cuda : "constexpr dim3" -> "const dim3"
5408d55506 cuda : uint -> uint32_t
f725ca90fb ggml : ggml_soft_max support F16/F32 mask/pos
c11d05fec0 llama : force disable flash attention for incompatible models
Compare 7 commits »
happyz synced commits to refs/pull/6757/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:01 -07:00
d27b4d3f2d Merge b79a41ee6a into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6766/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:01 -07:00
95f170a894 Merge c2691d968a into e931888d50
c2691d968a disable for multi-gpu and batch size > 1
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
800f4fe48e Tidied to now only use CUDA runtime (not mixed with driver calls)
Compare 8 commits »
happyz synced commits to refs/pull/6766/head at happyz/llama.cpp from mirror 2024-04-22 11:14:01 -07:00
c2691d968a disable for multi-gpu and batch size > 1
800f4fe48e Tidied to now only use CUDA runtime (not mixed with driver calls)
c8dd0e7c1c FIx issues raised in comments
Compare 3 commits »
happyz synced commits to refs/pull/6767/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:01 -07:00
7be551d40a Merge 84158931cf into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6721/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:00 -07:00
de53de7f7b Merge 8d4e096b09 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 16 commits »
happyz synced commits to refs/pull/6739/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:00 -07:00
ca9088b922 Merge ca0409fae4 into e931888d50
ca0409fae4 Merge branch 'ggerganov:master' into mannix-server-startup
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
Compare 6 commits »
happyz synced commits to refs/pull/6707/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:00 -07:00
b1cf3a1b0f Merge a43eb826eb into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6688/merge at happyz/llama.cpp from mirror 2024-04-22 11:14:00 -07:00
ae140b7a15 Merge 141eb5107f into e931888d50
141eb5107f Update llama_model_quantize_params
d6e453eb6c Split model correctly even if tensor id is out-of-order
6d66e609b5 Update examples/quantize/quantize.cpp
e931888d50 ggml : fix calloc argument ordering. (#6820)
Compare 17 commits »
happyz synced commits to refs/pull/6739/head at happyz/llama.cpp from mirror 2024-04-22 11:14:00 -07:00
ca0409fae4 Merge branch 'ggerganov:master' into mannix-server-startup
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 21 commits »
happyz synced commits to refs/pull/6648/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:59 -07:00
77276e4406 Merge ab2fae200c into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6658/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:59 -07:00
08d2a03aa1 Merge 8360e0c960 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6688/head at happyz/llama.cpp from mirror 2024-04-22 11:13:59 -07:00
141eb5107f Update llama_model_quantize_params
d6e453eb6c Split model correctly even if tensor id is out-of-order
6d66e609b5 Update examples/quantize/quantize.cpp
Compare 3 commits »
happyz synced commits to refs/pull/6640/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:59 -07:00
24771b8191 Merge a9a2983630 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6644/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:59 -07:00
4b69474d00 Merge 1b988855dca2ced3850dbe40812707e639b1dbd6 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 8 commits »
happyz synced commits to refs/pull/6511/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:58 -07:00
e072f62002 Merge a6f54dee3ca65ae1dbeab3f8c26c1d75a9609715 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6638/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:58 -07:00
5aeeeea17a Merge 6c257f4709 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6602/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:58 -07:00
2e16592dda Merge 1cd0a03720 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6563/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:58 -07:00
0745e03478 Merge 9acb43d7fa0b8da867570c975d33f0728951ca46 into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »
happyz synced commits to refs/pull/6445/merge at happyz/llama.cpp from mirror 2024-04-22 11:13:58 -07:00
2e50dcb1e7 Merge 91f3db8aab into e931888d50
e931888d50 ggml : fix calloc argument ordering. (#6820)
8960fe86ae llama : fix typo in <|im_end|> token text (#6745)
c0956b09ba ci: fix job are cancelling each other (#6781)
e9b4a1bf68 flake.lock: Update
Compare 5 commits »