HappyZ happyz
happyz synced commits to refs/pull/7931/head at happyz/llama.cpp from mirror 2024-06-14 09:23:36 -07:00
7a8961fff5 delete redundant
happyz synced commits to refs/pull/7915/head at happyz/llama.cpp from mirror 2024-06-14 09:23:35 -07:00
b30565e0c8 rpc : enable async operations
172c825684 rpc : fix ggml_backend_rpc_supports_buft() (#7918)
Compare 2 commits »
happyz synced commits to refs/pull/7921/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:35 -07:00
1d9dd480ff rever q2_K precision related changes
bff3a20944 fix data race
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
Compare 7 commits »
happyz synced commits to refs/pull/7921/head at happyz/llama.cpp from mirror 2024-06-14 09:23:35 -07:00
1d9dd480ff rever q2_K precision related changes
bff3a20944 fix data race
Compare 2 commits »
happyz synced commits to refs/pull/7919/head at happyz/llama.cpp from mirror 2024-06-14 09:23:35 -07:00
ded54b5d9b Replace powf with sycl::pow in ggml-sycl.cpp
happyz synced commits to refs/pull/7915/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:35 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 6 commits »
happyz synced commits to refs/pull/7896/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:34 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7899/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:34 -07:00
077e858a25 Merge bd28e1d22850b49614d51a6c7f9183e49ca55b20 into 66ef1ceedf
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7909/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:34 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7910/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:34 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7858/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:33 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7853/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:33 -07:00
9e7cb01325 Merge deafd1c7918c1873c824ff21e279053b7dd1035d into 66ef1ceedf
deafd1c791 gguf-dump: right align element count
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
a048c7fb6c gguf-dump.py: prettyfy dimention
e65bbf606c llama-bench : fix RPC indication (#7936)
Compare 8 commits »
happyz synced commits to refs/pull/7845/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:33 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7851/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:33 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7888/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:33 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
41b9260f18 convert : add Poro-34B-chat tokenizer support (#7713)
Compare 5 commits »
happyz synced commits to refs/pull/7853/head at happyz/llama.cpp from mirror 2024-06-14 09:23:33 -07:00
deafd1c791 gguf-dump: right align element count
a048c7fb6c gguf-dump.py: prettyfy dimention
9e2d2d917f Apply suggestions from code review
Compare 3 commits »
happyz synced commits to refs/pull/7843/head at happyz/llama.cpp from mirror 2024-06-14 09:23:32 -07:00
225ec48fe5 np.int16 no longer used
069369f3fe fix masking in __compute_fp32_to_bf16
Compare 2 commits »
happyz synced commits to refs/pull/7844/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:32 -07:00
181c0e3b0f review: modify codes as review suggestion
11c7b1e25a review: modify codes as review suggestion
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
Compare 7 commits »
happyz synced commits to refs/pull/7844/head at happyz/llama.cpp from mirror 2024-06-14 09:23:32 -07:00
181c0e3b0f review: modify codes as review suggestion
11c7b1e25a review: modify codes as review suggestion
Compare 2 commits »
happyz synced commits to refs/pull/7843/merge at happyz/llama.cpp from mirror 2024-06-14 09:23:32 -07:00
66ef1ceedf metal : utilize max shared memory for mul_mat_id (#7935)
e65bbf606c llama-bench : fix RPC indication (#7936)
225ec48fe5 np.int16 no longer used
6fcd1331ef llama : more checks before assuming FIM tokens (#7644)
Compare 7 commits »