162 lines
29 KiB
Plaintext
162 lines
29 KiB
Plaintext
fd9bd632f346085920d523dd307a3e4c11cc0a05 e08e473cf292ba73789769b993bb35bcdfc31689 wangbomeng wangbomeng@calculet.tech 2026-06-01T14:15:35+08:00 HEAD -> dev, origin/dev llama-bench: fix device-info
|
|
e08e473cf292ba73789769b993bb35bcdfc31689 f86112d6758298241ab2fa84ad4f0f7b1ace35c6 wangbomeng wangbomeng@calculet.tech 2026-06-01T11:47:01+08:00 llama-bench:add device-info
|
|
f86112d6758298241ab2fa84ad4f0f7b1ace35c6 d1bc743c10080fed1253686ce01749b42611ded4 wangbomeng wangbomeng@calculet.tech 2026-05-29T16:40:38+08:00 add dump_pos,fix pos_tensor of pre_by_embeds
|
|
d1bc743c10080fed1253686ce01749b42611ded4 2f201160b1960fd7be18b70d5770599d7082dd2f wangbomeng wangbomeng@calculet.tech 2026-05-28T14:23:06+08:00 set u_batch = max_seq_len
|
|
2f201160b1960fd7be18b70d5770599d7082dd2f f131bf1908b8c5dd1f49d7ae7f616506fc471666 wangbomeng wangbomeng@calculet.tech 2026-05-27T09:58:31+08:00 delete useless code
|
|
f131bf1908b8c5dd1f49d7ae7f616506fc471666 33d96dae28e17208ef61a71dba40cda3c04f135a wangbomeng wangbomeng@calculet.tech 2026-05-27T09:38:51+08:00 fix patch of qwen2.5vl,n_patch of minicpm = 64,fix n_ctx_slot,comment out M-ROPE
|
|
ebce9e1a548c2329aa97f53b8b2ad50d968088dc e850ff4de25ad1221abc9bf3bb30df252ccd9c36 Yunzhe Jia yunzhe@calculet.tech 2026-05-27T09:38:00+08:00 origin/runtime_replace Refactor cal-llm integration by removing runtime capacity loading and adjusting request handling
|
|
e850ff4de25ad1221abc9bf3bb30df252ccd9c36 9985688f7fa96f755882f96a2cd8a18114fdc42d Yunzhe Jia yunzhe@calculet.tech 2026-05-26T10:59:45+08:00 Enhance cal-llm integration by adding runtime checks for CALRT_LIBRARY and updating link settings for Linux and Windows
|
|
9985688f7fa96f755882f96a2cd8a18114fdc42d 38ca6895815671a87ceec86b4aa4eb976580ce76 Yunzhe Jia yunzhe@calculet.tech 2026-05-25T09:05:49+08:00 Refactor CalrtEngineConfig initialization by removing redundant parameters
|
|
33d96dae28e17208ef61a71dba40cda3c04f135a 37f068d87cc2d0a0b40ce978031fc9932fc89077 wangbomeng wangbomeng@calculet.tech 2026-05-22T14:33:45+08:00 add llama-build.sh
|
|
1921024604db640f09ade83c8a437ebf15bde360 d028722a12c5a463cc97207787cfd03d7a2590b0 5af229fe518ee2d761ab8ae2a69f7a9aaa9057f4 Bomeng Wang wangbomeng@calculet.tech 2026-05-21T14:08:14+08:00 refs/stash WIP on dev: d028722a replace cparam.n_ubatch with n_ubatch()
|
|
5af229fe518ee2d761ab8ae2a69f7a9aaa9057f4 d028722a12c5a463cc97207787cfd03d7a2590b0 Bomeng Wang wangbomeng@calculet.tech 2026-05-21T14:08:14+08:00 index on dev: d028722a replace cparam.n_ubatch with n_ubatch()
|
|
37f068d87cc2d0a0b40ce978031fc9932fc89077 13508e5e18bbb1202ae60c2ee5ecbdda1cca817b wangbomeng wangbomeng@calculet.tech 2026-05-21T14:07:00+08:00 release
|
|
13508e5e18bbb1202ae60c2ee5ecbdda1cca817b b6b7919468126d5ba30a10941825154789e9082f wangbomeng wangbomeng@calculet.tech 2026-05-21T11:43:59+08:00 reduce cmake log
|
|
b6b7919468126d5ba30a10941825154789e9082f 46847c24fc5c52dd57bb4fb17fc4d684149bb4fa wangbomeng wangbomeng@calculet.tech 2026-05-20T16:40:33+08:00 reduce cmake log
|
|
38ca6895815671a87ceec86b4aa4eb976580ce76 ddd60c928224b52eb4e1e60e01e8a1ecbb80ee45 Yunzhe Jia yunzhe@calculet.tech 2026-05-20T16:37:13+08:00 Add support for cal-llm profile dumping to JSONL file
|
|
46847c24fc5c52dd57bb4fb17fc4d684149bb4fa d028722a12c5a463cc97207787cfd03d7a2590b0 wangbomeng wangbomeng@calculet.tech 2026-05-20T16:02:12+08:00 Simplify CMakeLists.txt
|
|
d028722a12c5a463cc97207787cfd03d7a2590b0 63442fde8dc285817f725bd6c5fae80f478a3813 wangbomeng wangbomeng@calculet.tech 2026-05-20T14:41:05+08:00 replace cparam.n_ubatch with n_ubatch()
|
|
63442fde8dc285817f725bd6c5fae80f478a3813 fc3f4a9fab88e98eff8eeca8cbc6617333edba2c wangbomeng wangbomeng@calculet.tech 2026-05-20T11:47:42+08:00 fix ubatch and cmakelist
|
|
ddd60c928224b52eb4e1e60e01e8a1ecbb80ee45 2716130010c58f1cdd9d94f0e2f8e047bc92dc66 Yunzhe Jia yunzhe@calculet.tech 2026-05-20T09:22:40+08:00 Add cal-llm backend profiling support with command-line option
|
|
fc3f4a9fab88e98eff8eeca8cbc6617333edba2c c931ef3163940227c9b6291e04216245cce21429 wangbomeng wangbomeng@calculet.tech 2026-05-19T16:54:16+08:00 add dump and rt case generation
|
|
c931ef3163940227c9b6291e04216245cce21429 a43741244a646d3f8fa3ff01bb9b7d059223d0a5 wangbomeng wangbomeng@calculet.tech 2026-05-18T17:34:37+08:00 calrt: fix prefill_by_ids.fix create_buf
|
|
2716130010c58f1cdd9d94f0e2f8e047bc92dc66 b35bda124ccd4ed581dd6088d3ffa4931c2809b6 Yunzhe Jia yunzhe@calculet.tech 2026-05-18T08:40:54+08:00 fix n_ctx_slot error by adding runtime capacity loading for cal-llm
|
|
a43741244a646d3f8fa3ff01bb9b7d059223d0a5 b2003d6c90ef0fa26ab9e930ba3e22b355eb3d39 wangbomeng wangbomeng@calculet.tech 2026-05-15T14:39:53+08:00 run llama-passkey successfully, fix llama-bench, copy 3 times input to infer qwen2.5vl
|
|
b35bda124ccd4ed581dd6088d3ffa4931c2809b6 81a2cbd4cef5d28334ff513d6b69e593d9d1cb49 Yunzhe Jia yunzhe@calculet.tech 2026-05-15T09:19:31+08:00 Implement vocab-only model initialization and enhance cal-llm backend error handling
|
|
81a2cbd4cef5d28334ff513d6b69e593d9d1cb49 3b14ed556c15529fe1f94b649f49db7156dfaf67 Yunzhe Jia yunzhe@calculet.tech 2026-05-14T16:06:37+08:00 Enhance CMake configuration for cal-llm integration by updating paths and adding library properties
|
|
3b14ed556c15529fe1f94b649f49db7156dfaf67 dde8b8228081769c8d8f1e0751d47fa7875f8423 Yunzhe Jia yunzhe@calculet.tech 2026-05-14T14:34:25+08:00 - Removed references to CALRT from clip.h, mtmd-helper.cpp, mtmd-helper.h, mtmd.cpp, and mtmd.h. - Updated server.cpp to conditionally include CAL-LLM support. - Introduced option LLAMA_SERVER_USE_CAL_LLM in CMakeLists.txt to enable CAL-LLM integration. - Refactored server context to handle CAL-LLM requests and responses. - Updated image processing logic to remove CALRT-specific implementations. - Ensured compatibility with existing llama functionality while integrating new CAL-LLM features.
|
|
b2003d6c90ef0fa26ab9e930ba3e22b355eb3d39 7800c772d44a7415640ec1669a691232939927cf wangbomeng wangbomeng@calculet.tech 2026-05-13T15:24:24+08:00 try passkey and embedding
|
|
7800c772d44a7415640ec1669a691232939927cf dde8b8228081769c8d8f1e0751d47fa7875f8423 wangbomeng wangbomeng@calculet.tech 2026-05-13T15:23:54+08:00 try passkey and embedding
|
|
dde8b8228081769c8d8f1e0751d47fa7875f8423 fabcc7dac4b34d83fe430980bb364b81087c7ee9 wangbomeng wangbomeng@calculet.tech 2026-05-12T10:46:08+08:00 clip:add minicpm patch;
|
|
fabcc7dac4b34d83fe430980bb364b81087c7ee9 1e9787962f20a7e82c71589ac1dd0585454e4bfe wangbomeng wangbomeng@calculet.tech 2026-04-30T16:17:17+08:00 calrt: fix encode dump
|
|
1e9787962f20a7e82c71589ac1dd0585454e4bfe f584311c716dd078c4b737eebbfda97fe12b7274 1bec0db5a675ddc60fb793be2a746e8f3c8855dc wangbomeng wangbomeng@calculet.tech 2026-04-30T16:06:23+08:00 Merge remote-tracking branch 'origin/dev-ben' into dev
|
|
f584311c716dd078c4b737eebbfda97fe12b7274 a339954cc2a145a80c5ea349e7896211916cbed9 wangbomeng wangbomeng@calculet.tech 2026-04-30T16:05:15+08:00 calrt: fix encode dump
|
|
1bec0db5a675ddc60fb793be2a746e8f3c8855dc 465df20c3e1f3f58d0f5531c796e0941a49c32b0 wangbomeng wangbomeng@calculet.tech 2026-04-30T16:02:59+08:00 origin/dev-ben bench: support 2 calbins
|
|
465df20c3e1f3f58d0f5531c796e0941a49c32b0 a339954cc2a145a80c5ea349e7896211916cbed9 wangbomeng wangbomeng@calculet.tech 2026-04-30T09:08:12+08:00 support one calbin
|
|
a339954cc2a145a80c5ea349e7896211916cbed9 8e76fc330ad624d1768b412d7894e272dbe696f4 eeeef3f6d6a5aad9296323ca15d9d929745b8932 wangbomeng wangbomeng@calculet.tech 2026-04-29T14:57:55+08:00 tag: v0.3.0 Merge branch 'dev' into dev-cli
|
|
8e76fc330ad624d1768b412d7894e272dbe696f4 431ad6f1c787f1b47e9e7be14b2e1a7d8565485c wangbomeng wangbomeng@calculet.tech 2026-04-29T14:56:59+08:00 finish dev of llama-cli
|
|
eeeef3f6d6a5aad9296323ca15d9d929745b8932 431ad6f1c787f1b47e9e7be14b2e1a7d8565485c wangbomeng wangbomeng@calculet.tech 2026-04-28T16:16:19+08:00 add unknown time info: prefill and decode
|
|
431ad6f1c787f1b47e9e7be14b2e1a7d8565485c efa1876bb8b2a449a55054a8c3745892f2c0f262 wangbomeng wangbomeng@calculet.tech 2026-04-27T16:52:15+08:00 update calrt_infer
|
|
efa1876bb8b2a449a55054a8c3745892f2c0f262 4a2a3181f1cea170105c771e6ba69d851b82b937 wangbomeng wangbomeng@calculet.tech 2026-04-27T11:35:58+08:00 tag: v0.2.2 update version print
|
|
4a2a3181f1cea170105c771e6ba69d851b82b937 e3ea5541259f35871c3e995331770aaca297e9ec wangbomeng wangbomeng@calculet.tech 2026-04-26T10:56:02+08:00 add calrt-version
|
|
e3ea5541259f35871c3e995331770aaca297e9ec 0b8e43943846a1cd05c26c9186b9ef778ebae6e2 wangbomeng wangbomeng@calculet.tech 2026-04-26T10:55:48+08:00 add calrt-version
|
|
0b8e43943846a1cd05c26c9186b9ef778ebae6e2 5eb46fffbd2e260c540fb112314ad67bbb8f56b6 wangbomeng wangbomeng@calculet.tech 2026-04-26T08:43:43+08:00 to debug
|
|
5eb46fffbd2e260c540fb112314ad67bbb8f56b6 005112c51e00d998b457337ca3dad53fc46e2d0b wangbomeng wangbomeng@calculet.tech 2026-04-24T16:24:51+08:00 clip: add image_to_patches; calrt: vision copy data from bf16 to fp32
|
|
005112c51e00d998b457337ca3dad53fc46e2d0b 509670642e8090a1713f5d73590b549be827aaef wangbomeng wangbomeng@calculet.tech 2026-04-24T16:24:31+08:00 clip: add image_to_patches; calrt: vision copy data from bf16 to fp32
|
|
509670642e8090a1713f5d73590b549be827aaef 5d219bfcef239c1e3ddd1baecc0514a9462ff6a8 wangbomeng wangbomeng@calculet.tech 2026-04-24T09:40:34+08:00 rm debug code
|
|
5d219bfcef239c1e3ddd1baecc0514a9462ff6a8 6787a53ab9b4897c5a7b51a4062d7b90e6697d35 wangbomeng wangbomeng@calculet.tech 2026-04-24T09:31:32+08:00 calrt: recover base = batch_index * dict_len
|
|
6787a53ab9b4897c5a7b51a4062d7b90e6697d35 e1a76abf66178ea85656479a5e19aea7785c09f4 wangbomeng wangbomeng@calculet.tech 2026-04-24T09:22:02+08:00 calrt:input of multimodel from fp32 to bf16
|
|
e1a76abf66178ea85656479a5e19aea7785c09f4 99dd308adb8488fa869bb669e3335e646a938eff wangbomeng wangbomeng@calculet.tech 2026-04-22T16:29:22+08:00 calrt:add CalcoreRT version print
|
|
99dd308adb8488fa869bb669e3335e646a938eff 06c7852e99e34c71e24e3ef6001cb54c72970e4d wangbomeng wangbomeng@calculet.tech 2026-04-21T15:50:01+08:00 tag: v0.2.1 server: fix max_seq_len
|
|
06c7852e99e34c71e24e3ef6001cb54c72970e4d 96b4e267e562f44e76ec6c965ee0fa361f0439bf ed62eb33447f993d578e156a3c7e38c91b420ece wangbomeng wangbomeng@calculet.tech 2026-04-21T15:06:01+08:00 Merge branch 'dev-ot' into dev
|
|
ed62eb33447f993d578e156a3c7e38c91b420ece 015e988a271573a7c11e2a4b45084e1219d69d50 wangbomeng wangbomeng@calculet.tech 2026-04-21T15:05:22+08:00 try server-test:not success
|
|
015e988a271573a7c11e2a4b45084e1219d69d50 99962021f2dd1994dbbf90d5712f0ca2633f20f8 wangbomeng wangbomeng@calculet.tech 2026-04-17T16:56:08+08:00 little change
|
|
96b4e267e562f44e76ec6c965ee0fa361f0439bf 99962021f2dd1994dbbf90d5712f0ca2633f20f8 wangbomeng wangbomeng@calculet.tech 2026-04-17T15:33:01+08:00 merge dev-ot
|
|
99962021f2dd1994dbbf90d5712f0ca2633f20f8 9af6682ef101c6d006c743c703cf150ac25f466f wangbomeng wangbomeng@calculet.tech 2026-04-17T14:48:33+08:00 origin/dev-ot calrt: add onetoken slice and untile_prefill
|
|
9af6682ef101c6d006c743c703cf150ac25f466f e24f775ee26ea5249d5ff6cd23b8a0a17fe1438d wangbomeng wangbomeng@calculet.tech 2026-04-16T10:09:44+08:00 args: add -ca
|
|
e24f775ee26ea5249d5ff6cd23b8a0a17fe1438d a206460e610aef995e36e50b33e8bd41fe477541 wangbomeng wangbomeng@calculet.tech 2026-04-15T10:01:09+08:00 calrt: add copy_data_to_ibuf for vision model
|
|
795179beb5bba728e0ccf0a7e80a5ec5ef36e92a 6d55eae7e76b768acfcec045b8e84cded8ae3b55 Yunzhe Jia yunzhe@calculet.tech 2026-04-14T16:03:22+08:00 origin/test-cicd add ci yaml
|
|
a206460e610aef995e36e50b33e8bd41fe477541 d8b2aaa00049978cebac15041f960d75b552f97b wangbomeng wangbomeng@calculet.tech 2026-04-14T10:23:43+08:00 calrt: support mixture of prefill_by_ids and prefill_by_embeds
|
|
d8b2aaa00049978cebac15041f960d75b552f97b b6f29ec8140028619efaa50aa6e73146cecf631d wangbomeng wangbomeng@calculet.tech 2026-04-13T11:23:44+08:00 server: fix context-shift
|
|
b6f29ec8140028619efaa50aa6e73146cecf631d 6d55eae7e76b768acfcec045b8e84cded8ae3b55 wangbomeng wangbomeng@calculet.tech 2026-04-13T11:23:31+08:00 server: fix context-shift
|
|
6d55eae7e76b768acfcec045b8e84cded8ae3b55 af8055f5f033a730e7e1fe80185b18b7e9473966 wangbomeng wangbomeng@calculet.tech 2026-04-10T14:58:56+08:00 tag: v0.2.0 add annotations
|
|
af8055f5f033a730e7e1fe80185b18b7e9473966 6f54682df62b8ec6bb4154b99c47b5becda3291f wangbomeng wangbomeng@calculet.tech 2026-04-10T14:58:19+08:00 add annotations
|
|
6f54682df62b8ec6bb4154b99c47b5becda3291f 42a181674565411efa5c8b318bc77d6fd084df8a a2d5bf362d6c9a546f9deadf4f8975605a915d33 Yunzhe Jia yunzhe@calculet.tech 2026-04-10T09:07:20+08:00 Merge branch 'dev' of http://192.168.10.26:1000/yunzhe/llama.cpp into dev
|
|
42a181674565411efa5c8b318bc77d6fd084df8a bd3a57a11aa008e0b5dc2a30f82fcd3a02587208 Yunzhe Jia yunzhe@calculet.tech 2026-04-10T09:06:55+08:00 fix prompt_cache issue
|
|
a2d5bf362d6c9a546f9deadf4f8975605a915d33 bd3a57a11aa008e0b5dc2a30f82fcd3a02587208 Bomeng Wang wangbomeng@calculet.tech 2026-04-09T21:01:46+08:00 server: comment fixed time
|
|
bd3a57a11aa008e0b5dc2a30f82fcd3a02587208 fcfcdf80b01664a08fb12df879a05df3cc0eba70 74d437ab1456831573def7ca385a74d0ca460c37 Yunzhe Jia yunzhe@calculet.tech 2026-04-09T19:42:57+08:00 Merge branch 'dev' of http://192.168.10.26:1000/yunzhe/llama.cpp into dev
|
|
fcfcdf80b01664a08fb12df879a05df3cc0eba70 0cd44929b442860dc0e5f88976108be82350f9dc Yunzhe Jia yunzhe@calculet.tech 2026-04-09T19:41:52+08:00 tmp fix one run >4K problem
|
|
74d437ab1456831573def7ca385a74d0ca460c37 5813c943005d1ede551e67ee98e35aefd210ff22 0cd44929b442860dc0e5f88976108be82350f9dc wangbomeng wangbomeng@calculet.tech 2026-04-09T17:44:26+08:00 Merge branch 'dev' of http://192.168.10.26:1000/yunzhe/llama.cpp into dev
|
|
5813c943005d1ede551e67ee98e35aefd210ff22 1f0d5bb0453f93ffa51faf68237c89251c2e0c40 wangbomeng wangbomeng@calculet.tech 2026-04-09T17:38:57+08:00 add --device-info
|
|
1f0d5bb0453f93ffa51faf68237c89251c2e0c40 b8153b6b8f244b223c9f4c2940ae1f212634519e wangbomeng wangbomeng@calculet.tech 2026-04-09T17:38:37+08:00 add --device-info
|
|
0cd44929b442860dc0e5f88976108be82350f9dc b8153b6b8f244b223c9f4c2940ae1f212634519e Bomeng Wang wangbomeng@calculet.tech 2026-04-09T15:08:50+08:00 rm calculet0.txt
|
|
b8153b6b8f244b223c9f4c2940ae1f212634519e 82842cfc3fa3d5c004c4540aeff8a81f38509bee wangbomeng wangbomeng@calculet.tech 2026-04-09T15:07:02+08:00 server: fix limit tokens to max_seq_len
|
|
82842cfc3fa3d5c004c4540aeff8a81f38509bee 622a22991089c6debd5f4b57738f3df3a0bda8b5 wangbomeng wangbomeng@calculet.tech 2026-04-09T15:06:26+08:00 server: fix limit tokens to max_seq_len
|
|
622a22991089c6debd5f4b57738f3df3a0bda8b5 5f47a738ba56a37d0ff281a16212f41979568e35 wangbomeng wangbomeng@calculet.tech 2026-04-09T10:36:31+08:00 server: comment circle print
|
|
5f47a738ba56a37d0ff281a16212f41979568e35 ecf01a17651cd7b1e8b85fd6075e83831463ee23 207ca38c751f7e6c251833eb11e89a2a0bb0fd07 Yunzhe Jia yunzhe@calculet.tech 2026-04-09T09:48:20+08:00 Merge branch 'dev' of http://192.168.10.26:1000/yunzhe/llama.cpp into dev
|
|
ecf01a17651cd7b1e8b85fd6075e83831463ee23 99d2078e89305035d87d966cee7987c7d50b3925 Yunzhe Jia yunzhe@calculet.tech 2026-04-09T09:47:55+08:00 fix n_parallel problem
|
|
207ca38c751f7e6c251833eb11e89a2a0bb0fd07 99d2078e89305035d87d966cee7987c7d50b3925 wangbomeng wangbomeng@calculet.tech 2026-04-08T19:14:37+08:00 delet extra printf
|
|
99d2078e89305035d87d966cee7987c7d50b3925 9626239f5a9e568c9975d67dd6745fef90279a87 Yunzhe Jia yunzhe@calculet.tech 2026-04-08T19:01:16+08:00 fix buffer resize problem
|
|
9626239f5a9e568c9975d67dd6745fef90279a87 da724f3e2a8af2c2537e3edcff41e9415ac66fef wangbomeng wangbomeng@calculet.tech 2026-04-08T17:47:04+08:00 calrt: fix MapBuf
|
|
da724f3e2a8af2c2537e3edcff41e9415ac66fef 18814f9bf1a423ebe8f32d6043c9cbc5b21b0218 Bomeng Wang wangbomeng@calculet.tech 2026-04-08T16:58:32+08:00 calrt: add MapBuf
|
|
18814f9bf1a423ebe8f32d6043c9cbc5b21b0218 bbee06d6d851e3f84f75b578047967d7e0e4a8dc wangbomeng wangbomeng@calculet.tech 2026-04-08T13:55:52+08:00 server : reset inference time
|
|
bbee06d6d851e3f84f75b578047967d7e0e4a8dc e126c21ed7b793b648d63533ea7ee30885f5a82a wangbomeng wangbomeng@calculet.tech 2026-04-08T11:29:36+08:00 calrt: add test time print
|
|
39edc6ec2e9560ccb20a2b8e547f58102bb268c1 2006831fd6ea8a8e16518b81bcee29fb674676ea Bomeng Wang wangbomeng@calculet.tech 2026-04-01T14:06:50+08:00 tag: v0.1.1 add t_incopy t_outcopy
|
|
2006831fd6ea8a8e16518b81bcee29fb674676ea 41e6636ea1fb3884e41fb2023f92b0e7262ef09b wangbomeng wangbomeng@calculet.tech 2026-04-01T10:42:27+08:00 tag: v0.1.0 v0.1.0
|
|
e126c21ed7b793b648d63533ea7ee30885f5a82a 5603e050af07173589f3b8850121263d86da01fd wangbomeng wangbomeng@calculet.tech 2026-04-01T10:07:12+08:00 sever: fix commit id
|
|
5603e050af07173589f3b8850121263d86da01fd 1e8a16cd0e3ad3c9a0f8746852a0a38c6e52322e wangbomeng wangbomeng@calculet.tech 2026-04-01T05:57:55+08:00 calrt: add t_incopy and t_outcopy
|
|
1e8a16cd0e3ad3c9a0f8746852a0a38c6e52322e 893e52bda0fee99bbda1521ea733a760bc4201f1 wangbomeng wangbomeng@calculet.tech 2026-04-01T10:14:22+08:00 calrt: fix slice and add infer/copy time
|
|
893e52bda0fee99bbda1521ea733a760bc4201f1 398ecf71f5f574037ce33f84dd1576fa164909ad Bomeng Wang wangbomeng@calculet.tech 2026-03-30T11:35:40+08:00 delete unnecessary commit id
|
|
398ecf71f5f574037ce33f84dd1576fa164909ad 6d5e83902bdd3142a8ff0d3db19f02a8619e95a0 wangbomeng wangbomeng@calculet.tech 2026-03-27T02:01:23+08:00 rm llama-server_old.cpp
|
|
6d5e83902bdd3142a8ff0d3db19f02a8619e95a0 e4eaab700333490e80ebefc8d023f6c1adfcc312 wangbomeng wangbomeng@calculet.tech 2026-03-27T00:57:11+08:00 server: printf calrt commit ID
|
|
e4eaab700333490e80ebefc8d023f6c1adfcc312 aeacc23a883400db0ebffb50e23355ada054da66 wangbomeng wangbomeng@calculet.tech 2026-03-25T09:03:55+08:00 calrt: prefil to prefill
|
|
aeacc23a883400db0ebffb50e23355ada054da66 cbaa7492663630132adeddccd9fec5e44ffa3336 wangbomeng wangbomeng@calculet.tech 2026-03-25T08:49:22+08:00 calrt: add slice only on decode
|
|
cbaa7492663630132adeddccd9fec5e44ffa3336 3c08ebf0e442cd730ddffd959ec676f9caf31c82 wangbomeng wangbomeng@calculet.tech 2026-03-25T03:21:02+08:00 reduce warning
|
|
3c08ebf0e442cd730ddffd959ec676f9caf31c82 613bc2b9a9d7af92efbe42d876ffcdb244f7a00e wangbomeng wangbomeng@calculet.tech 2026-03-24T02:43:03+08:00 calrt: comment LOG_ERROR in untile_decode_cpy
|
|
613bc2b9a9d7af92efbe42d876ffcdb244f7a00e 4e194ecc6c4b8ebe4c72d93f6055e3605dbfcefa wangbomeng wangbomeng@calculet.tech 2026-03-24T02:30:12+08:00 calrt: fix max_seq_len judgment
|
|
4e194ecc6c4b8ebe4c72d93f6055e3605dbfcefa 16cbf81a731f5839a2f6b0eb9d8539826353748b wangbomeng wangbomeng@calculet.tech 2026-03-24T01:59:03+08:00 calrt: fix ubatch.embd cpy
|
|
16cbf81a731f5839a2f6b0eb9d8539826353748b 8b4e17d990c23025148c4abadefe1010a0dd0734 wangbomeng wangbomeng@calculet.tech 2026-03-23T09:38:21+08:00 calrt: fix model_name of untile_cpy
|
|
8b4e17d990c23025148c4abadefe1010a0dd0734 6dacdd2cf9574214ca272a86674e7bd7ba21c4f0 41e6636ea1fb3884e41fb2023f92b0e7262ef09b wangbomeng wangbomeng@calculet.tech 2026-03-23T14:50:27+08:00 merge origin dev
|
|
6dacdd2cf9574214ca272a86674e7bd7ba21c4f0 3e6ad64467771224f072e5bad90e270951ba4cdf wangbomeng wangbomeng@calculet.tech 2026-03-23T14:42:28+08:00 calrt: add multimodal
|
|
3e6ad64467771224f072e5bad90e270951ba4cdf b37b7d9756feb3dac2db526eeab38b572a095d47 wangbomeng wangbomeng@calculet.tech 2026-03-23T14:36:41+08:00 calrt: add multimodal
|
|
41e6636ea1fb3884e41fb2023f92b0e7262ef09b d0a1b2c758440720d42fa108b3444f54593daa73 Bomeng Wang wangbomeng@calculet.tech 2026-03-12T16:36:07+08:00 tag: v0.0.1 server: correct the max_seq_len judgment
|
|
d0a1b2c758440720d42fa108b3444f54593daa73 b37b7d9756feb3dac2db526eeab38b572a095d47 Bomeng Wang wangbomeng@calculet.tech 2026-03-12T16:10:28+08:00 calrt: comment out dump
|
|
b37b7d9756feb3dac2db526eeab38b572a095d47 07817cefc14fa359dcecad0630defd3610f8f5c8 7ddafbdcb9426ae3af016925b8b596f11e8742d0 Yunzhe Jia yunzhe@calculet.tech 2026-03-10T16:32:31+08:00 Merge branch 'dev' of http://192.168.10.26:1000/yunzhe/llama.cpp into dev
|
|
07817cefc14fa359dcecad0630defd3610f8f5c8 b2c45d8ddefb7ec09435231c8ec5b736fabf9c91 Yunzhe Jia yunzhe@calculet.tech 2026-03-10T16:32:07+08:00 fix kv issue
|
|
7ddafbdcb9426ae3af016925b8b596f11e8742d0 059009eeaf20a569abfbac5eb37dad3bb9376428 wangbomeng wangbomeng@calculet.tech 2026-03-10T07:13:04+08:00 calrt: commented out cal_ctx->encode(batch)
|
|
059009eeaf20a569abfbac5eb37dad3bb9376428 1b073661d1311b5300173993b9f31ee7241d8606 wangbomeng wangbomeng@calculet.tech 2026-03-10T07:04:22+08:00 calrt: add multimodel
|
|
1b073661d1311b5300173993b9f31ee7241d8606 6295273c5866b9249bacdedbfc3a31877e6e6a85 b2c45d8ddefb7ec09435231c8ec5b736fabf9c91 wangbomeng wangbomeng@calculet.tech 2026-03-10T06:58:35+08:00 Merge branch 'dev' of http://192.168.20.229:1000/yunzhe/llama.cpp into dev
|
|
6295273c5866b9249bacdedbfc3a31877e6e6a85 c6ea3f0d2468e3b8eab5de80bacee7fe4c9dd3d7 wangbomeng wangbomeng@calculet.tech 2026-03-10T06:54:41+08:00 arg: update hf_repo url to https://download.calculet.tech:9443
|
|
b2c45d8ddefb7ec09435231c8ec5b736fabf9c91 c6ea3f0d2468e3b8eab5de80bacee7fe4c9dd3d7 Bomeng Wang wangbomeng@calculet.tech 2026-03-09T15:03:26+08:00 clart: past_kv_len = n_tokens
|
|
c6ea3f0d2468e3b8eab5de80bacee7fe4c9dd3d7 f8fba66b657c51a1df1efc7267bfea9f6140cebf wangbomeng wangbomeng@calculet.tech 2026-03-09T03:26:47+08:00 CMakeLists: STATIC to SHARED
|
|
f8fba66b657c51a1df1efc7267bfea9f6140cebf db4a61d2234c2f457b42ecc3f7219a1e1a35508a Yunzhe Jia yunzhe@calculet.tech 2026-03-06T15:06:36+08:00 fix calrt_install include path
|
|
db4a61d2234c2f457b42ecc3f7219a1e1a35508a e672e02ca1a582f9e3a9d011d9e1b4fea4942a25 Yunzhe Jia yunzhe@calculet.tech 2026-03-06T06:59:58+08:00 adapt to host-rt install structure
|
|
e672e02ca1a582f9e3a9d011d9e1b4fea4942a25 27b82f64252779ad160a3fa76bd378c39421de36 7a3a9a0e76bb5f530beafcfe52e87e388e93117a Yunzhe Jia yunzhe@calculet.tech 2026-03-06T11:47:09+08:00 Merge branch 'dev' of http://192.168.10.26:1000/yunzhe/llama.cpp into dev
|
|
27b82f64252779ad160a3fa76bd378c39421de36 0bd95719fed2f0247c8d0199dd897d24157dfae6 Yunzhe Jia yunzhe@calculet.tech 2026-03-06T11:42:45+08:00 adapt to calrt_objs target
|
|
7a3a9a0e76bb5f530beafcfe52e87e388e93117a 583c387efaf7bc030574e554088de79d09a27fbd wangbomeng wangbomeng@calculet.tech 2026-03-05T07:24:59+08:00 calrt_infer: fix past_kv_len
|
|
583c387efaf7bc030574e554088de79d09a27fbd 0edc14a60a2d5596f71d728c655e3523da924d89 wangbomeng wangbomeng@calculet.tech 2026-02-27T02:24:56+08:00 caserver: lim generated tokens to n_ctx_per_seq
|
|
0edc14a60a2d5596f71d728c655e3523da924d89 aa0a17d719326a63e0d6fea60aa0ced44e7abc33 wangbomeng wangbomeng@calculet.tech 2026-02-27T01:57:39+08:00 caserver: limit generated tokens to n_ctx_per_seq
|
|
aa0a17d719326a63e0d6fea60aa0ced44e7abc33 fa332e773da681f3e77d6ffe83a87edb557f758b wangbomeng wangbomeng@calculet.tech 2026-02-27T01:40:11+08:00 calrt: add bf16_to_fp32
|
|
fa332e773da681f3e77d6ffe83a87edb557f758b aa54a7c637f2c92d924a5f35386975b1c27de898 wangbomeng wangbomeng@calculet.tech 2026-02-27T01:31:42+08:00 Add bf16_to_fp32
|
|
aa54a7c637f2c92d924a5f35386975b1c27de898 84f2c0b5f334a7ab77347c6779b6027a951d0ff0 wangbomeng wangbomeng@calculet.tech 2026-02-27T01:30:36+08:00 Add bf16_to_fp32
|
|
84f2c0b5f334a7ab77347c6779b6027a951d0ff0 0bd95719fed2f0247c8d0199dd897d24157dfae6 wangbomeng wangbomeng@calculet.tech 2026-02-09T01:46:36+08:00 add bf16_to_fp32 in untile_decode/prefill_cpy
|
|
0bd95719fed2f0247c8d0199dd897d24157dfae6 365eb0b719e30b0f9e348d854f11c9c3f954a96e Yunzhe Jia yunzhe@calculet.tech 2026-02-07T10:25:06+08:00 fix get_model_by_name
|
|
365eb0b719e30b0f9e348d854f11c9c3f954a96e 886cd1fb3b217efcf51c46209fcb694022346797 Yunzhe Jia yunzhe@calculet.tech 2026-02-07T09:05:17+08:00 comment out pos buffer file
|
|
886cd1fb3b217efcf51c46209fcb694022346797 9328f21378275f2354a19c3af907332b216ab492 Yunzhe Jia yunzhe@calculet.tech 2026-02-06T17:36:54+08:00 fix compile problem
|
|
9328f21378275f2354a19c3af907332b216ab492 49fea70a20f550bfdb59d0ffe041b54919c22447 Yunzhe Jia yunzhe@calculet.tech 2026-01-28T15:32:36+08:00 add prefill-only mode
|
|
49fea70a20f550bfdb59d0ffe041b54919c22447 74a118df31b6a4acc8876b2a2db90d2395cc6ec3 Yunzhe Jia yunzhe@calculet.tech 2026-01-28T09:11:16+08:00 adapt to new calbin test_case
|
|
74a118df31b6a4acc8876b2a2db90d2395cc6ec3 98e8f9acfe87cd93646482cb2d942b8a132f0852 Yunzhe Jia yunzhe@calculet.tech 2026-01-13T14:32:47+08:00 sync with calrt update
|
|
98e8f9acfe87cd93646482cb2d942b8a132f0852 8c8152a04036bb0d4756a4e08e54f514dae4e64d Yunzhe Jia yunzhe@calculet.tech 2025-12-10T00:20:57+08:00 add timestamp
|
|
8c8152a04036bb0d4756a4e08e54f514dae4e64d 5d2387931528fbe04aa8d66de935147efb65ca4d Yunzhe Jia yunzhe@calculet.tech 2025-12-09T14:40:04+08:00 adapt to calrt csr
|
|
5d2387931528fbe04aa8d66de935147efb65ca4d e6285a748657add8d974b78ca920c5b41753ee12 Yunzhe Jia yunzhe@calculet.tech 2025-12-08T16:40:34+08:00 add timestamp for one decode process
|
|
e6285a748657add8d974b78ca920c5b41753ee12 6b1ecdbbe3cc72ed8888be38f9e934cfd79b9fbb Yunzhe Jia yunzhe@calculet.tech 2025-11-27T16:58:35+08:00 adapt to calrt
|
|
6b1ecdbbe3cc72ed8888be38f9e934cfd79b9fbb 240d9bb75241c91896afe125f2fc74698a7395f3 Yunzhe Jia yunzhe@calculet.tech 2025-11-25T17:06:14+08:00 add kv_tensor for pld test
|
|
240d9bb75241c91896afe125f2fc74698a7395f3 7eeb241aad429a68fb8cfa3ba85d000fb2348cd2 Yunzhe Jia yunzhe@calculet.tech 2025-11-24T12:03:00+08:00 adapt to calrt modification
|
|
7eeb241aad429a68fb8cfa3ba85d000fb2348cd2 dfb6250031a25058dbc3757e8a7ec75d90dcb48a Yunzhe Jia yunzhe@calculet.tech 2025-11-21T09:20:01+08:00 fix byte_size error
|
|
dfb6250031a25058dbc3757e8a7ec75d90dcb48a e31bce25e85dc7fdf0a0e69fd0e58ba4a2ffaf3e Yunzhe Jia yunzhe@calculet.tech 2025-11-18T10:28:00+08:00 add elf in CMakelist
|
|
e31bce25e85dc7fdf0a0e69fd0e58ba4a2ffaf3e 6444e227c36df16199c40d8cead2244fe014f377 Yunzhe Jia yunzhe@calculet.tech 2025-11-07T11:52:17+08:00 dump src tensor
|
|
6444e227c36df16199c40d8cead2244fe014f377 d81d7b190ff5f21325167936496dfe5ab48b922e Yunzhe Jia yunzhe@calculet.tech 2025-11-07T09:19:11+08:00 test golden case
|
|
d81d7b190ff5f21325167936496dfe5ab48b922e e492f5dc8cec17529f53ef2bbdf7b502107f8148 1f5accb8d0056e6099cd5b772b1cb787dd590a13 Yunzhe Jia yunzhe@calculet.tech 2025-11-04T14:43:32+08:00 Merge branch 'master' into dev
|
|
e492f5dc8cec17529f53ef2bbdf7b502107f8148 2e05afd22b3d77c7013b11eee88e88b17188f21a Yunzhe Jia yunzhe@calculet.tech 2025-11-04T14:09:59+08:00 redo untile logic
|
|
2e05afd22b3d77c7013b11eee88e88b17188f21a c20c0a5c0c44179f5267a262546624854b1203d1 Yunzhe Jia yunzhe@calculet.tech 2025-10-31T14:06:06+08:00 add pcie support
|
|
c20c0a5c0c44179f5267a262546624854b1203d1 62425b3a7e8b169a3bbe6ce8c7324513a7fcee8c Yunzhe Jia yunzhe@calculet.tech 2025-10-20T14:41:08+08:00 kv: turn on common-prefix mtmd: fix output size bug
|
|
62425b3a7e8b169a3bbe6ce8c7324513a7fcee8c 76a63a2998286c1649de2496bccb7d8a46549895 Yunzhe Jia yunzhe@calculet.tech 2025-10-16T17:34:39+08:00 kv: let seq_rm free whole seq since current hw-op cannot support common-prefix
|
|
76a63a2998286c1649de2496bccb7d8a46549895 b67217078aa748b06e1b2269d88dd18c0aca2c26 Yunzhe Jia yunzhe@calculet.tech 2025-10-16T16:31:15+08:00 kv: fix common-prefix bug by implementing llama_memory_seq_rm
|
|
b67217078aa748b06e1b2269d88dd18c0aca2c26 98c8269abefdfa8fde47dfb2769ceebc8b704e86 Yunzhe Jia yunzhe@calculet.tech 2025-10-13T11:48:14+08:00 add different path for unified and seperate compilation
|
|
98c8269abefdfa8fde47dfb2769ceebc8b704e86 2790ecc304e384d9075c55d0628d8bf440025dc4 Yunzhe Jia yunzhe@calculet.tech 2025-09-30T17:42:57+08:00 server: adapt to calrt
|
|
2790ecc304e384d9075c55d0628d8bf440025dc4 f528e97d894b4d93934a0711a00253eea980e799 Yunzhe Jia yunzhe@calculet.tech 2025-09-30T17:37:58+08:00 adapt to calrt
|
|
f528e97d894b4d93934a0711a00253eea980e799 dcbe998a6d8f13b6c9cb36bbe7b355ced2ac3842 Yunzhe Jia yunzhe@calculet.tech 2025-09-11T14:13:35+08:00 add calrt mtmd support
|
|
dcbe998a6d8f13b6c9cb36bbe7b355ced2ac3842 87932ec016f0ec81e98a13ce5212851f8c12458a Yunzhe Jia yunzhe@calculet.tech 2025-09-05T14:34:07+08:00 add calrt API used in server side
|
|
87932ec016f0ec81e98a13ce5212851f8c12458a 0115cfb7c1dec0753aecb34f3af4338bf9de8eeb Yunzhe Jia yunzhe@calculet.tech 2025-08-27T14:29:18+08:00 fix merge problem
|
|
0115cfb7c1dec0753aecb34f3af4338bf9de8eeb 3a617cb18121a85eed17ac5585788810c5d9c1e0 0d5a470223fc90b6b6807921d68011ff06ae7f9e Yunzhe Jia yunzhe@calculet.tech 2025-08-26T15:19:56+08:00 Merge branch 'master' into dev
|
|
3a617cb18121a85eed17ac5585788810c5d9c1e0 c42293848dd9f03eba9a6c1b85a600b398f06541 Yunzhe Jia yunzhe@calculet.tech 2025-08-13T10:58:59+08:00 add calrt_decode
|
|
c42293848dd9f03eba9a6c1b85a600b398f06541 1cbf442f50a154aac0bf35d7b8b3436a2cdfd6f5 Yunzhe Jia yunzhe@calculet.tech 2025-08-06T15:46:06+08:00 sync with calrt API
|
|
1cbf442f50a154aac0bf35d7b8b3436a2cdfd6f5 ba67d6c03fed3193987a4ef13050e6e2a4451df4 Yunzhe Jia yunzhe@calculet.tech 2025-07-24T15:50:19+08:00 1.sync with calrt prefill,decode,kv_update op 2.add hw pattern untile function
|
|
ba67d6c03fed3193987a4ef13050e6e2a4451df4 8f52c6f8abb2ed9ad967cc0bfb90acb073011193 Yunzhe Jia yunzhe@calculet.tech 2025-06-26T17:42:32+08:00 llama-calrt: infer-op: copy data from output buffer to dst-tensor.data
|
|
8f52c6f8abb2ed9ad967cc0bfb90acb073011193 d2f325130c4ed77a3dcff2c23a926587fb08e2cc Yunzhe Jia yunzhe@calculet.tech 2025-06-26T15:10:45+08:00 Squashed commit of the following:
|
|
d2f325130c4ed77a3dcff2c23a926587fb08e2cc 6391ee7ffd97e723dfbdcc32f001e1b95a63cba6 Yunzhe Jia yunzhe@calculet.tech 2025-06-17T11:53:33+08:00 calrt: add input_token in graph; input_token will be copy to calrt_in_buffer
|
|
6391ee7ffd97e723dfbdcc32f001e1b95a63cba6 b9850050a5b5147cb2e6720b98ed8c1e8e02eff9 Yunzhe Jia yunzhe@calculet.tech 2025-06-13T10:58:06+08:00 update .gitignore to update cal_test.gguf
|
|
b9850050a5b5147cb2e6720b98ed8c1e8e02eff9 933b205c32551848589b8339437d778cb4f2a46d Yunzhe Jia yunzhe@calculet.tech 2025-06-06T17:25:52+08:00 calrt: add call chain for using calrt::infer
|
|
933b205c32551848589b8339437d778cb4f2a46d 33fea9a8592fd338abe279b89350a3af2e0ff1f7 Yunzhe Jia yunzhe@calculet.tech 2025-06-12T17:47:19+08:00 gguf-py: add cal_test.py
|
|
33fea9a8592fd338abe279b89350a3af2e0ff1f7 806f05a353d322e97d97e02f910de2b408a8b4ab Yunzhe Jia yunzhe@calculet.tech 2025-06-06T17:25:52+08:00 calrt: add load calbin
|
|
806f05a353d322e97d97e02f910de2b408a8b4ab ec9e0301fef6476df83e94842c3b625501c95566 Yunzhe Jia yunzhe@calculet.tech 2025-05-30T09:45:37+08:00 add calrt demo |