From fdfdfedeb44bd28551279a94dc972b4d4aed3b7d Mon Sep 17 00:00:00 2001 From: jhqxxx Date: Mon, 22 Sep 2025 23:49:58 +0800 Subject: [PATCH] update qwen2.5vl --- Cargo.lock | 967 ++++++++++++++++++++++++++++++- Cargo.toml | 7 +- src/lib.rs | 2 +- src/models/mod.rs | 7 +- src/models/qwen2_5vl/generate.rs | 163 ++++-- src/utils/utils.rs | 172 ++++++ tests/test_qwen2_5vl.rs | 56 +- 7 files changed, 1300 insertions(+), 74 deletions(-) diff --git a/Cargo.lock b/Cargo.lock index 0ce3965..65de170 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -27,15 +27,19 @@ dependencies = [ "candle-flash-attn", "candle-nn", "candle-transformers", + "chrono", "ffmpeg-next", "image", "minijinja", "num", "openai_dive", "reqwest", + "rocket", "serde", "serde_json", "tokenizers", + "tokio", + "uuid", ] [[package]] @@ -70,6 +74,15 @@ dependencies = [ "equator", ] +[[package]] +name = "android_system_properties" +version = "0.1.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "819e7219dbd41043ac279b19830f2efc897156490d7fd6ea916720117ee66311" +dependencies = [ + "libc", +] + [[package]] name = "anyhow" version = "1.0.100" @@ -105,6 +118,54 @@ version = "0.7.6" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "7c02d123df017efcdfbd739ef81735b36c5ba83ec3c59c80a9d7ecc718f92e50" +[[package]] +name = "async-stream" +version = "0.3.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b5a71a6f37880a80d1d7f19efd781e4b5de42c88f0722cc13bcb6cc2cfe8476" +dependencies = [ + "async-stream-impl", + "futures-core", + "pin-project-lite", +] + +[[package]] +name = "async-stream-impl" +version = "0.3.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c7c24de15d275a1ecfd47a380fb4d5ec9bfe0933f309ed5e705b775596a3574d" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "async-trait" +version = "0.1.89" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9035ad2d096bed7955a320ee7e2230574d28fd3c3a0f186cbea1ff3c7eed5dbb" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "atomic" +version = "0.5.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c59bdb34bc650a32731b31bd8f0829cc15d24a708ee31559e0bb34f2bc320cba" + +[[package]] +name = "atomic" +version = "0.6.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a89cbf775b137e9b968e67227ef7f775587cde3fd31b0d8599dbd0f598a48340" +dependencies = [ + "bytemuck", +] + [[package]] name = "atomic-waker" version = "1.1.2" @@ -167,6 +228,12 @@ version = "0.22.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "72b3254f16251a8381aa12e40e3c4d2f0199f8c6508fbecb9d91f575e0fbb8c6" +[[package]] +name = "binascii" +version = "0.1.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "383d29d513d8764dcdc42ea295d979eb99c3c9f00607b3692cf68a431f7dca72" + [[package]] name = "bindgen" version = "0.72.1" @@ -411,6 +478,19 @@ version = "1.0.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "2fd1289c04a9ea8cb22300a459a72a385d7c73d3259e2ed7dcb2af674838cfa9" +[[package]] +name = "chrono" +version = "0.4.42" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "145052bdd345b87320e369255277e3fb5152762ad123a901ef5c262dd38fe8d2" +dependencies = [ + "iana-time-zone", + "js-sys", + "num-traits", + "wasm-bindgen", + "windows-link 0.2.0", +] + [[package]] name = "clang-sys" version = "1.8.1" @@ -456,6 +536,17 @@ dependencies = [ "windows-sys 0.59.0", ] +[[package]] +name = "cookie" +version = "0.18.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4ddef33a339a91ea89fb53151bd0a4689cfce27055c291dfa69945475d22c747" +dependencies = [ + "percent-encoding", + "time", + "version_check", +] + [[package]] name = "core-foundation" version = "0.9.4" @@ -567,6 +658,15 @@ dependencies = [ "serde", ] +[[package]] +name = "deranged" +version = "0.5.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d630bccd429a5bb5a64b5e94f693bfc48c9f8566418fda4c494cc94f911f87cc" +dependencies = [ + "powerfmt", +] + [[package]] name = "derive_arbitrary" version = "1.4.2" @@ -609,6 +709,39 @@ dependencies = [ "syn", ] +[[package]] +name = "devise" +version = "0.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f1d90b0c4c777a2cad215e3c7be59ac7c15adf45cf76317009b7d096d46f651d" +dependencies = [ + "devise_codegen", + "devise_core", +] + +[[package]] +name = "devise_codegen" +version = "0.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "71b28680d8be17a570a2334922518be6adc3f58ecc880cbb404eaeb8624fd867" +dependencies = [ + "devise_core", + "quote", +] + +[[package]] +name = "devise_core" +version = "0.4.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b035a542cf7abf01f2e3c4d5a7acbaebfefe120ae4efc7bde3df98186e4b8af7" +dependencies = [ + "bitflags 2.9.4", + "proc-macro2", + "proc-macro2-diagnostics", + "quote", + "syn", +] + [[package]] name = "displaydoc" version = "0.2.5" @@ -717,6 +850,17 @@ dependencies = [ "cc", ] +[[package]] +name = "eventsource-stream" +version = "0.2.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "74fef4569247a5f429d9156b9d0a2599914385dd189c539334c625d8099d90ab" +dependencies = [ + "futures-core", + "nom", + "pin-project-lite", +] + [[package]] name = "exr" version = "1.73.0" @@ -803,6 +947,20 @@ dependencies = [ "vcpkg", ] +[[package]] +name = "figment" +version = "0.10.19" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8cb01cd46b0cf372153850f4c6c272d9cbea2da513e07538405148f95bd789f3" +dependencies = [ + "atomic 0.6.1", + "pear", + "serde", + "toml", + "uncased", + "version_check", +] + [[package]] name = "find-msvc-tools" version = "0.1.2" @@ -871,6 +1029,21 @@ dependencies = [ "percent-encoding", ] +[[package]] +name = "futures" +version = "0.3.31" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "65bc07b1a8bc7c85c5f2e110c476c7389b4554ba72af57d8445ea63a576b0876" +dependencies = [ + "futures-channel", + "futures-core", + "futures-executor", + "futures-io", + "futures-sink", + "futures-task", + "futures-util", +] + [[package]] name = "futures-channel" version = "0.3.31" @@ -887,6 +1060,17 @@ version = "0.3.31" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "05f29059c0c2090612e8d742178b0580d2dc940c837851ad723096f87af6663e" +[[package]] +name = "futures-executor" +version = "0.3.31" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1e28d1d997f585e54aebc3f97d39e72338912123a67330d723fdbb564d646c9f" +dependencies = [ + "futures-core", + "futures-task", + "futures-util", +] + [[package]] name = "futures-io" version = "0.3.31" @@ -916,12 +1100,19 @@ version = "0.3.31" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "f90f7dce0722e95104fcb095585910c0977252f286e354b5e3bd38902cd99988" +[[package]] +name = "futures-timer" +version = "3.0.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f288b0a4f20f9a56b5d1da57e2227c661b7b16168e2f72365f57b63326e29b24" + [[package]] name = "futures-util" version = "0.3.31" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9fa08315bb612088cc391249efdc3bc77536f16c91f6cf495e6fbe85b20a4a81" dependencies = [ + "futures-channel", "futures-core", "futures-io", "futures-macro", @@ -1170,6 +1361,19 @@ dependencies = [ "seq-macro", ] +[[package]] +name = "generator" +version = "0.7.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5cc16584ff22b460a382b7feec54b23d2908d858152e5739a120b949293bd74e" +dependencies = [ + "cc", + "libc", + "log", + "rustversion", + "windows", +] + [[package]] name = "getrandom" version = "0.2.16" @@ -1215,6 +1419,25 @@ version = "0.3.3" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0cc23270f6e1808e30a928bdc84dea0b9b4136a8bc82338574f23baf47bbd280" +[[package]] +name = "h2" +version = "0.3.27" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0beca50380b1fc32983fc1cb4587bfa4bb9e78fc259aad4a0032d2080309222d" +dependencies = [ + "bytes", + "fnv", + "futures-core", + "futures-sink", + "futures-util", + "http 0.2.12", + "indexmap", + "slab", + "tokio", + "tokio-util", + "tracing", +] + [[package]] name = "h2" version = "0.4.12" @@ -1226,7 +1449,7 @@ dependencies = [ "fnv", "futures-core", "futures-sink", - "http", + "http 1.3.1", "indexmap", "slab", "tokio", @@ -1266,6 +1489,17 @@ version = "0.5.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "fc0fef456e4baa96da950455cd02c081ca953b141298e41db3fc7e36b1da849c" +[[package]] +name = "http" +version = "0.2.12" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "601cbb57e577e2f5ef5be8e7b83f0f63994f25aa94d673e54a92d5c516d101f1" +dependencies = [ + "bytes", + "fnv", + "itoa", +] + [[package]] name = "http" version = "1.3.1" @@ -1277,6 +1511,17 @@ dependencies = [ "itoa", ] +[[package]] +name = "http-body" +version = "0.4.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7ceab25649e9960c0311ea418d17bee82c0dcec1bd053b5f9a66e265a693bed2" +dependencies = [ + "bytes", + "http 0.2.12", + "pin-project-lite", +] + [[package]] name = "http-body" version = "1.0.1" @@ -1284,7 +1529,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "1efedce1fb8e6913f23e0c92de8e62cd5b772a67e7b3946df930a62566c93184" dependencies = [ "bytes", - "http", + "http 1.3.1", ] [[package]] @@ -1295,8 +1540,8 @@ checksum = "b021d93e26becf5dc7e1b75b1bed1fd93124b374ceb73f43d4d4eafec896a64a" dependencies = [ "bytes", "futures-core", - "http", - "http-body", + "http 1.3.1", + "http-body 1.0.1", "pin-project-lite", ] @@ -1306,6 +1551,36 @@ version = "1.10.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "6dbf3de79e51f3d586ab4cb9d5c3e2c14aa28ed23d180cf89b4df0454a69cc87" +[[package]] +name = "httpdate" +version = "1.0.3" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "df3b46402a9d5adb4c86a0cf463f42e19994e3ee891101b1841f30a545cb49a9" + +[[package]] +name = "hyper" +version = "0.14.32" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "41dfc780fdec9373c01bae43289ea34c972e40ee3c9f6b3c8801a35f35586ce7" +dependencies = [ + "bytes", + "futures-channel", + "futures-core", + "futures-util", + "h2 0.3.27", + "http 0.2.12", + "http-body 0.4.6", + "httparse", + "httpdate", + "itoa", + "pin-project-lite", + "socket2 0.5.10", + "tokio", + "tower-service", + "tracing", + "want", +] + [[package]] name = "hyper" version = "1.7.0" @@ -1316,9 +1591,9 @@ dependencies = [ "bytes", "futures-channel", "futures-core", - "h2", - "http", - "http-body", + "h2 0.4.12", + "http 1.3.1", + "http-body 1.0.1", "httparse", "itoa", "pin-project-lite", @@ -1334,8 +1609,8 @@ version = "0.27.7" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "e3c93eb611681b207e1fe55d5a71ecf91572ec8a6705cdb6857f7d8d5242cf58" dependencies = [ - "http", - "hyper", + "http 1.3.1", + "hyper 1.7.0", "hyper-util", "rustls", "rustls-pki-types", @@ -1352,7 +1627,7 @@ checksum = "70206fc6890eaca9fde8a0bf71caa2ddfc9fe045ac9e5c70df101a7dbde866e0" dependencies = [ "bytes", "http-body-util", - "hyper", + "hyper 1.7.0", "hyper-util", "native-tls", "tokio", @@ -1371,14 +1646,14 @@ dependencies = [ "futures-channel", "futures-core", "futures-util", - "http", - "http-body", - "hyper", + "http 1.3.1", + "http-body 1.0.1", + "hyper 1.7.0", "ipnet", "libc", "percent-encoding", "pin-project-lite", - "socket2", + "socket2 0.6.0", "system-configuration", "tokio", "tower-service", @@ -1386,6 +1661,30 @@ dependencies = [ "windows-registry", ] +[[package]] +name = "iana-time-zone" +version = "0.1.64" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "33e57f83510bb73707521ebaffa789ec8caf86f9657cad665b092b581d40e9fb" +dependencies = [ + "android_system_properties", + "core-foundation-sys", + "iana-time-zone-haiku", + "js-sys", + "log", + "wasm-bindgen", + "windows-core", +] + +[[package]] +name = "iana-time-zone-haiku" +version = "0.1.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f31827a206f56af32e590ba56d5d2d085f558508192593743f16b2306495269f" +dependencies = [ + "cc", +] + [[package]] name = "icu_collections" version = "2.0.0" @@ -1547,6 +1846,8 @@ checksum = "4b0f83760fb341a774ed326568e19f5a863af4a952def8c39f9ab92fd95b88e5" dependencies = [ "equivalent", "hashbrown", + "serde", + "serde_core", ] [[package]] @@ -1562,6 +1863,12 @@ dependencies = [ "web-time", ] +[[package]] +name = "inlinable_string" +version = "0.1.15" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "c8fae54786f62fb2918dcfae3d568594e50eb9b5c25bf04371af6fe7516452fb" + [[package]] name = "interpolate_name" version = "0.2.4" @@ -1600,6 +1907,17 @@ dependencies = [ "serde", ] +[[package]] +name = "is-terminal" +version = "0.4.16" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e04d7f318608d35d4b61ddd75cbdaee86b023ebe2bd5a66ee0915f0bf93095a9" +dependencies = [ + "hermit-abi", + "libc", + "windows-sys 0.59.0", +] + [[package]] name = "itertools" version = "0.12.1" @@ -1653,6 +1971,12 @@ dependencies = [ "wasm-bindgen", ] +[[package]] +name = "lazy_static" +version = "1.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bbd2bcb4c963f2ddae06a2efc7e9f3591312473c50c6685e1f298068316e66fe" + [[package]] name = "lebe" version = "0.5.3" @@ -1703,12 +2027,37 @@ version = "0.8.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "241eaef5fd12c88705a01fc1066c48c4b36e0dd4377dcdc7ec3942cea7a69956" +[[package]] +name = "lock_api" +version = "0.4.13" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "96936507f153605bddfcda068dd804796c84324ed2510809e5b2a624c81da765" +dependencies = [ + "autocfg", + "scopeguard", +] + [[package]] name = "log" version = "0.4.28" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "34080505efa8e45a4b816c349525ebe327ceaa8559756f0356cba97ef3bf7432" +[[package]] +name = "loom" +version = "0.5.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ff50ecb28bb86013e935fb6683ab1f6d3a20016f123c76fd4c27470076ac30f5" +dependencies = [ + "cfg-if", + "generator", + "scoped-tls", + "serde", + "serde_json", + "tracing", + "tracing-subscriber", +] + [[package]] name = "loop9" version = "0.1.5" @@ -1734,6 +2083,15 @@ version = "0.2.2" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "670fdfda89751bc4a84ac13eaa63e205cf0fd22b4c9a5fbfa085b63c1f1d3a30" +[[package]] +name = "matchers" +version = "0.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d1525a2a28c7f4fa0fc98bb91ae755d1e2d1505079e05539e35bc876b5d65ae9" +dependencies = [ + "regex-automata", +] + [[package]] name = "maybe-rayon" version = "0.1.1" @@ -1844,6 +2202,25 @@ dependencies = [ "pxfm", ] +[[package]] +name = "multer" +version = "3.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "83e87776546dc87511aa5ee218730c92b666d7264ab6ed41f9d215af9cd5224b" +dependencies = [ + "bytes", + "encoding_rs", + "futures-util", + "http 1.3.1", + "httparse", + "memchr", + "mime", + "spin", + "tokio", + "tokio-util", + "version_check", +] + [[package]] name = "native-tls" version = "0.2.14" @@ -1883,6 +2260,15 @@ version = "0.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0676bb32a98c1a483ce53e500a81ad9c3d5b3f7c920c28c24e9cb0980d0b5bc8" +[[package]] +name = "nu-ansi-term" +version = "0.50.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "d4a28e057d01f97e61255210fcff094d74ed0466038633e95017f5beb68e4399" +dependencies = [ + "windows-sys 0.52.0", +] + [[package]] name = "num" version = "0.4.3" @@ -1917,6 +2303,12 @@ dependencies = [ "num-traits", ] +[[package]] +name = "num-conv" +version = "0.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "51d515d32fb182ee37cda2ccdcb92950d6a3c2893aa280e540671c2cd0f3b1d9" + [[package]] name = "num-derive" version = "0.4.2" @@ -2052,11 +2444,14 @@ checksum = "bac82ff995cdbe4e120fa87fe4d76bd452866dd7b0c9bcbdb562ba4be0f1786d" dependencies = [ "bytes", "derive_builder", + "futures", "reqwest", + "reqwest-eventsource", "serde", "serde_html_form", "serde_json", "tokio", + "tokio-stream", "tokio-util", ] @@ -2104,12 +2499,58 @@ dependencies = [ "vcpkg", ] +[[package]] +name = "parking_lot" +version = "0.12.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "70d58bf43669b5795d1576d0641cfb6fbb2057bf629506267a92807158584a13" +dependencies = [ + "lock_api", + "parking_lot_core", +] + +[[package]] +name = "parking_lot_core" +version = "0.9.11" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bc838d2a56b5b1a6c25f55575dfc605fabb63bb2365f6c2353ef9159aa69e4a5" +dependencies = [ + "cfg-if", + "libc", + "redox_syscall", + "smallvec", + "windows-targets 0.52.6", +] + [[package]] name = "paste" version = "1.0.15" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "57c0d7b74b563b49d38dae00a0c37d4d6de9b432382b2892f0574ddcae73fd0a" +[[package]] +name = "pear" +version = "0.2.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bdeeaa00ce488657faba8ebf44ab9361f9365a97bd39ffb8a60663f57ff4b467" +dependencies = [ + "inlinable_string", + "pear_codegen", + "yansi", +] + +[[package]] +name = "pear_codegen" +version = "0.2.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4bab5b985dc082b345f812b7df84e1bef27e7207b39e448439ba8bd69c93f147" +dependencies = [ + "proc-macro2", + "proc-macro2-diagnostics", + "quote", + "syn", +] + [[package]] name = "percent-encoding" version = "2.3.2" @@ -2162,6 +2603,12 @@ dependencies = [ "zerovec", ] +[[package]] +name = "powerfmt" +version = "0.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "439ee305def115ba05938db6eb1644ff94165c5ab5e9420d1c1bcedbba909391" + [[package]] name = "ppv-lite86" version = "0.2.21" @@ -2189,6 +2636,19 @@ dependencies = [ "unicode-ident", ] +[[package]] +name = "proc-macro2-diagnostics" +version = "0.10.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "af066a9c399a26e020ada66a034357a868728e72cd426f3adcd35f80d88d88c8" +dependencies = [ + "proc-macro2", + "quote", + "syn", + "version_check", + "yansi", +] + [[package]] name = "profiling" version = "1.0.17" @@ -2447,6 +2907,35 @@ version = "0.5.5" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "03251193000f4bd3b042892be858ee50e8b3719f2b08e5833ac4353724632430" +[[package]] +name = "redox_syscall" +version = "0.5.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5407465600fb0548f1442edf71dd20683c6ed326200ace4b1ef0763521bb3b77" +dependencies = [ + "bitflags 2.9.4", +] + +[[package]] +name = "ref-cast" +version = "1.0.24" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4a0ae411dbe946a674d89546582cea4ba2bb8defac896622d6496f14c23ba5cf" +dependencies = [ + "ref-cast-impl", +] + +[[package]] +name = "ref-cast-impl" +version = "1.0.24" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "1165225c21bff1f3bbce98f5a1f889949bc902d3575308cc7b0de30b4f6d27c7" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + [[package]] name = "regex" version = "1.11.2" @@ -2488,11 +2977,11 @@ dependencies = [ "futures-channel", "futures-core", "futures-util", - "h2", - "http", - "http-body", + "h2 0.4.12", + "http 1.3.1", + "http-body 1.0.1", "http-body-util", - "hyper", + "hyper 1.7.0", "hyper-rustls", "hyper-tls", "hyper-util", @@ -2521,6 +3010,22 @@ dependencies = [ "web-sys", ] +[[package]] +name = "reqwest-eventsource" +version = "0.6.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "632c55746dbb44275691640e7b40c907c16a2dc1a5842aa98aaec90da6ec6bde" +dependencies = [ + "eventsource-stream", + "futures-core", + "futures-timer", + "mime", + "nom", + "pin-project-lite", + "reqwest", + "thiserror 1.0.69", +] + [[package]] name = "rgb" version = "0.8.52" @@ -2541,6 +3046,87 @@ dependencies = [ "windows-sys 0.52.0", ] +[[package]] +name = "rocket" +version = "0.5.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a516907296a31df7dc04310e7043b61d71954d703b603cc6867a026d7e72d73f" +dependencies = [ + "async-stream", + "async-trait", + "atomic 0.5.3", + "binascii", + "bytes", + "either", + "figment", + "futures", + "indexmap", + "log", + "memchr", + "multer", + "num_cpus", + "parking_lot", + "pin-project-lite", + "rand 0.8.5", + "ref-cast", + "rocket_codegen", + "rocket_http", + "serde", + "state", + "tempfile", + "time", + "tokio", + "tokio-stream", + "tokio-util", + "ubyte", + "version_check", + "yansi", +] + +[[package]] +name = "rocket_codegen" +version = "0.5.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "575d32d7ec1a9770108c879fc7c47815a80073f96ca07ff9525a94fcede1dd46" +dependencies = [ + "devise", + "glob", + "indexmap", + "proc-macro2", + "quote", + "rocket_http", + "syn", + "unicode-xid", + "version_check", +] + +[[package]] +name = "rocket_http" +version = "0.5.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e274915a20ee3065f611c044bd63c40757396b6dbc057d6046aec27f14f882b9" +dependencies = [ + "cookie", + "either", + "futures", + "http 0.2.12", + "hyper 0.14.32", + "indexmap", + "log", + "memchr", + "pear", + "percent-encoding", + "pin-project-lite", + "ref-cast", + "serde", + "smallvec", + "stable-pattern", + "state", + "time", + "tokio", + "uncased", +] + [[package]] name = "rustc-demangle" version = "0.1.26" @@ -2639,6 +3225,18 @@ dependencies = [ "windows-sys 0.61.0", ] +[[package]] +name = "scoped-tls" +version = "1.0.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e1cf6437eb19a8f4a6cc0f7dca544973b0b78843adbfeb3683d1a94a0024a294" + +[[package]] +name = "scopeguard" +version = "1.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49" + [[package]] name = "security-framework" version = "2.11.1" @@ -2754,12 +3352,30 @@ dependencies = [ "serde", ] +[[package]] +name = "sharded-slab" +version = "0.1.7" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f40ca3c46823713e0d4209592e8d6e826aa57e928f09752619fc696c499637f6" +dependencies = [ + "lazy_static", +] + [[package]] name = "shlex" version = "1.3.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64" +[[package]] +name = "signal-hook-registry" +version = "1.4.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "b2a4719bff48cee6b39d12c020eeb490953ad2443b7055bd0b21fca26bd8c28b" +dependencies = [ + "libc", +] + [[package]] name = "simd-adler32" version = "0.3.7" @@ -2787,6 +3403,16 @@ version = "1.15.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "67b1b7a3b5fe4f1376887184045fcf45c69e92af734b7aaddc05fb777b6fbd03" +[[package]] +name = "socket2" +version = "0.5.10" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e22376abed350d73dd1cd119b57ffccad95b4e585a7cda43e286245ce23c0678" +dependencies = [ + "libc", + "windows-sys 0.52.0", +] + [[package]] name = "socket2" version = "0.6.0" @@ -2797,6 +3423,12 @@ dependencies = [ "windows-sys 0.59.0", ] +[[package]] +name = "spin" +version = "0.9.8" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6980e8d7511241f8acf4aebddbb1ff938df5eebe98691418c4468d0b72a96a67" + [[package]] name = "spm_precompiled" version = "0.1.4" @@ -2809,12 +3441,30 @@ dependencies = [ "unicode-segmentation", ] +[[package]] +name = "stable-pattern" +version = "0.1.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "4564168c00635f88eaed410d5efa8131afa8d8699a612c80c455a0ba05c21045" +dependencies = [ + "memchr", +] + [[package]] name = "stable_deref_trait" version = "1.2.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "a8f112729512f8e442d81f95a8a7ddf2b7c6b8a1a6f509a95864142b30cab2d3" +[[package]] +name = "state" +version = "0.6.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2b8c4a4445d81357df8b1a650d0d0d6fbbbfe99d064aa5e02f3e4022061476d8" +dependencies = [ + "loom", +] + [[package]] name = "static_assertions" version = "1.1.0" @@ -2985,6 +3635,15 @@ dependencies = [ "syn", ] +[[package]] +name = "thread_local" +version = "1.1.9" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f60246a4944f24f6e018aa17cdeffb7818b76356965d03b07d6a9886e8962185" +dependencies = [ + "cfg-if", +] + [[package]] name = "tiff" version = "0.10.3" @@ -2999,6 +3658,37 @@ dependencies = [ "zune-jpeg", ] +[[package]] +name = "time" +version = "0.3.44" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "91e7d9e3bb61134e77bde20dd4825b97c010155709965fedf0f49bb138e52a9d" +dependencies = [ + "deranged", + "itoa", + "num-conv", + "powerfmt", + "serde", + "time-core", + "time-macros", +] + +[[package]] +name = "time-core" +version = "0.1.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "40868e7c1d2f0b8d73e4a8c7f0ff63af4f6d19be117e90bd73eb1d62cf831c6b" + +[[package]] +name = "time-macros" +version = "0.2.24" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "30cfb0125f12d9c277f35663a0a33f8c30190f4e4574868a330595412d34ebf3" +dependencies = [ + "num-conv", + "time-core", +] + [[package]] name = "tinystr" version = "0.8.1" @@ -3055,11 +3745,24 @@ dependencies = [ "libc", "mio", "pin-project-lite", + "signal-hook-registry", "slab", - "socket2", + "socket2 0.6.0", + "tokio-macros", "windows-sys 0.59.0", ] +[[package]] +name = "tokio-macros" +version = "2.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "6e06d43f1345a3bcd39f6a56dbb7dcab2ba47e68e8ac134855e7e2bdbaf8cab8" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + [[package]] name = "tokio-native-tls" version = "0.3.1" @@ -3080,6 +3783,17 @@ dependencies = [ "tokio", ] +[[package]] +name = "tokio-stream" +version = "0.1.17" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "eca58d7bba4a75707817a2c44174253f9236b2d5fbd055602e9d5c07c139a047" +dependencies = [ + "futures-core", + "pin-project-lite", + "tokio", +] + [[package]] name = "tokio-util" version = "0.7.16" @@ -3133,6 +3847,7 @@ dependencies = [ "serde", "serde_spanned", "toml_datetime 0.6.11", + "toml_write", "winnow", ] @@ -3157,6 +3872,12 @@ dependencies = [ "winnow", ] +[[package]] +name = "toml_write" +version = "0.1.2" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "5d99f8c9a7727884afe522e9bd5edbfc91a3312b36a77b5fb8926e4c31a41801" + [[package]] name = "tower" version = "0.5.2" @@ -3181,8 +3902,8 @@ dependencies = [ "bitflags 2.9.4", "bytes", "futures-util", - "http", - "http-body", + "http 1.3.1", + "http-body 1.0.1", "iri-string", "pin-project-lite", "tower", @@ -3231,6 +3952,36 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "b9d12581f227e93f094d3af2ae690a574abb8a2b9b7a96e7cfe9647b2b617678" dependencies = [ "once_cell", + "valuable", +] + +[[package]] +name = "tracing-log" +version = "0.2.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ee855f1f400bd0e5c02d150ae5de3840039a3f54b025156404e34c23c03f47c3" +dependencies = [ + "log", + "once_cell", + "tracing-core", +] + +[[package]] +name = "tracing-subscriber" +version = "0.3.20" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2054a14f5307d601f88daf0553e1cbf472acc4f2c51afab632431cdcd72124d5" +dependencies = [ + "matchers", + "nu-ansi-term", + "once_cell", + "regex-automata", + "sharded-slab", + "smallvec", + "thread_local", + "tracing", + "tracing-core", + "tracing-log", ] [[package]] @@ -3239,6 +3990,15 @@ version = "0.2.5" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "e421abadd41a4225275504ea4d6566923418b7f05506fbc9c0fe86ba7396114b" +[[package]] +name = "ubyte" +version = "0.10.4" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "f720def6ce1ee2fc44d40ac9ed6d3a59c361c80a75a7aa8e75bb9baed31cf2ea" +dependencies = [ + "serde", +] + [[package]] name = "ug" version = "0.5.0" @@ -3273,6 +4033,16 @@ dependencies = [ "ug", ] +[[package]] +name = "uncased" +version = "0.9.10" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e1b88fcfe09e89d3866a5c11019378088af2d24c3fbd4f0543f96b479ec90697" +dependencies = [ + "serde", + "version_check", +] + [[package]] name = "unicase" version = "2.8.1" @@ -3306,6 +4076,12 @@ version = "0.2.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "4a1a07cc7db3810833284e8d372ccdc6da29741639ecc70c9ec107df0fa6154c" +[[package]] +name = "unicode-xid" +version = "0.2.6" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ebc1c04c71510c7f702b52b7c350734c9ff1295c464a03335b00bb84fc54f853" + [[package]] name = "unicode_categories" version = "0.1.1" @@ -3336,6 +4112,17 @@ version = "1.0.4" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "b6c140620e7ffbb22c2dee59cafe6084a59b5ffc27a8859a5f0d494b5d52b6be" +[[package]] +name = "uuid" +version = "1.18.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2f87b8aa10b915a06587d0dec516c282ff295b475d94abf425d62b57710070a2" +dependencies = [ + "getrandom 0.3.3", + "js-sys", + "wasm-bindgen", +] + [[package]] name = "v_frame" version = "0.3.9" @@ -3347,6 +4134,12 @@ dependencies = [ "wasm-bindgen", ] +[[package]] +name = "valuable" +version = "0.1.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ba73ea9cf16a25df0c8caa16c51acb937d5712a8429db78a3ee29d5dcacd3a65" + [[package]] name = "vcpkg" version = "0.2.15" @@ -3528,6 +4321,50 @@ dependencies = [ "windows-sys 0.61.0", ] +[[package]] +name = "windows" +version = "0.48.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "e686886bc078bc1b0b600cac0147aadb815089b6e4da64016cbd754b6342700f" +dependencies = [ + "windows-targets 0.48.5", +] + +[[package]] +name = "windows-core" +version = "0.62.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "57fe7168f7de578d2d8a05b07fd61870d2e73b4020e9f49aa00da8471723497c" +dependencies = [ + "windows-implement", + "windows-interface", + "windows-link 0.2.0", + "windows-result 0.4.0", + "windows-strings 0.5.0", +] + +[[package]] +name = "windows-implement" +version = "0.60.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a47fddd13af08290e67f4acabf4b459f647552718f683a7b415d290ac744a836" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + +[[package]] +name = "windows-interface" +version = "0.59.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "bd9211b69f8dcdfa817bfd14bf1c97c9188afa36f4750130fcdf3f400eca9fa8" +dependencies = [ + "proc-macro2", + "quote", + "syn", +] + [[package]] name = "windows-link" version = "0.1.3" @@ -3547,8 +4384,8 @@ source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "5b8a9ed28765efc97bbc954883f4e6796c33a06546ebafacbabee9696967499e" dependencies = [ "windows-link 0.1.3", - "windows-result", - "windows-strings", + "windows-result 0.3.4", + "windows-strings 0.4.2", ] [[package]] @@ -3560,6 +4397,15 @@ dependencies = [ "windows-link 0.1.3", ] +[[package]] +name = "windows-result" +version = "0.4.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7084dcc306f89883455a206237404d3eaf961e5bd7e0f312f7c91f57eb44167f" +dependencies = [ + "windows-link 0.2.0", +] + [[package]] name = "windows-strings" version = "0.4.2" @@ -3569,6 +4415,15 @@ dependencies = [ "windows-link 0.1.3", ] +[[package]] +name = "windows-strings" +version = "0.5.0" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "7218c655a553b0bed4426cf54b20d7ba363ef543b52d515b3e48d7fd55318dda" +dependencies = [ + "windows-link 0.2.0", +] + [[package]] name = "windows-sys" version = "0.52.0" @@ -3596,6 +4451,21 @@ dependencies = [ "windows-link 0.2.0", ] +[[package]] +name = "windows-targets" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "9a2fa6e2155d7247be68c096456083145c183cbbbc2764150dda45a87197940c" +dependencies = [ + "windows_aarch64_gnullvm 0.48.5", + "windows_aarch64_msvc 0.48.5", + "windows_i686_gnu 0.48.5", + "windows_i686_msvc 0.48.5", + "windows_x86_64_gnu 0.48.5", + "windows_x86_64_gnullvm 0.48.5", + "windows_x86_64_msvc 0.48.5", +] + [[package]] name = "windows-targets" version = "0.52.6" @@ -3629,6 +4499,12 @@ dependencies = [ "windows_x86_64_msvc 0.53.0", ] +[[package]] +name = "windows_aarch64_gnullvm" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "2b38e32f0abccf9987a4e3079dfb67dcd799fb61361e53e2882c3cbaf0d905d8" + [[package]] name = "windows_aarch64_gnullvm" version = "0.52.6" @@ -3641,6 +4517,12 @@ version = "0.53.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "86b8d5f90ddd19cb4a147a5fa63ca848db3df085e25fee3cc10b39b6eebae764" +[[package]] +name = "windows_aarch64_msvc" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "dc35310971f3b2dbbf3f0690a219f40e2d9afcf64f9ab7cc1be722937c26b4bc" + [[package]] name = "windows_aarch64_msvc" version = "0.52.6" @@ -3653,6 +4535,12 @@ version = "0.53.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "c7651a1f62a11b8cbd5e0d42526e55f2c99886c77e007179efff86c2b137e66c" +[[package]] +name = "windows_i686_gnu" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "a75915e7def60c94dcef72200b9a8e58e5091744960da64ec734a6c6e9b3743e" + [[package]] name = "windows_i686_gnu" version = "0.52.6" @@ -3677,6 +4565,12 @@ version = "0.53.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "9ce6ccbdedbf6d6354471319e781c0dfef054c81fbc7cf83f338a4296c0cae11" +[[package]] +name = "windows_i686_msvc" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "8f55c233f70c4b27f66c523580f78f1004e8b5a8b659e05a4eb49d4166cca406" + [[package]] name = "windows_i686_msvc" version = "0.52.6" @@ -3689,6 +4583,12 @@ version = "0.53.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "581fee95406bb13382d2f65cd4a908ca7b1e4c2f1917f143ba16efe98a589b5d" +[[package]] +name = "windows_x86_64_gnu" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "53d40abd2583d23e4718fddf1ebec84dbff8381c07cae67ff7768bbf19c6718e" + [[package]] name = "windows_x86_64_gnu" version = "0.52.6" @@ -3701,6 +4601,12 @@ version = "0.53.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "2e55b5ac9ea33f2fc1716d1742db15574fd6fc8dadc51caab1c16a3d3b4190ba" +[[package]] +name = "windows_x86_64_gnullvm" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "0b7b52767868a23d5bab768e390dc5f5c55825b6d30b86c844ff2dc7414044cc" + [[package]] name = "windows_x86_64_gnullvm" version = "0.52.6" @@ -3713,6 +4619,12 @@ version = "0.53.0" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "0a6e035dd0599267ce1ee132e51c27dd29437f63325753051e71dd9e42406c57" +[[package]] +name = "windows_x86_64_msvc" +version = "0.48.5" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "ed94fce61571a4006852b7389a063ab983c02eb1bb37b47f8272ce92d06d9538" + [[package]] name = "windows_x86_64_msvc" version = "0.52.6" @@ -3746,6 +4658,15 @@ version = "0.6.1" source = "registry+https://github.com/rust-lang/crates.io-index" checksum = "ea2f10b9bb0928dfb1b42b65e1f9e36f7f54dbdf08457afefb38afcdec4fa2bb" +[[package]] +name = "yansi" +version = "1.0.1" +source = "registry+https://github.com/rust-lang/crates.io-index" +checksum = "cfe53a6657fd280eaa890a3bc59152892ffa3e30101319d168b781ed6529b049" +dependencies = [ + "is-terminal", +] + [[package]] name = "yoke" version = "0.7.5" diff --git a/Cargo.toml b/Cargo.toml index bc67317..c4c6dd4 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -21,7 +21,12 @@ base64 = "0.22.1" num = "0.4.3" minijinja = "2.12.0" tokenizers = "0.22.1" -openai_dive = "1.3.0" +openai_dive = { version = "1.3.0", features = ["stream"]} +uuid = { version = "1.18.1", features = ["v4"]} +chrono = "0.4.42" +rocket = "0.5.1" +tokio = "1.47.1" + [features] flash-attn=["candle-flash-attn"] cuda=["candle-nn/cuda", "candle-core/cuda", "candle-transformers/cuda"] diff --git a/src/lib.rs b/src/lib.rs index 897c0a9..fb19131 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -17,7 +17,7 @@ impl ModelType { model_path: &str, device: Option<&Device>, dtype: Option, - ) -> Result> { + ) -> Result> { match model_type { ModelType::Qwen2_5VL => { let model = Qwen2_5VLGenerateModel::init(model_path, device, dtype)?; diff --git a/src/models/mod.rs b/src/models/mod.rs index 10140c7..e66b18e 100644 --- a/src/models/mod.rs +++ b/src/models/mod.rs @@ -1,11 +1,14 @@ pub mod qwen2_5vl; use anyhow::Result; use candle_core::{DType, Device}; -use openai_dive::v1::resources::chat::ChatCompletionParameters; +use openai_dive::v1::resources::chat::{ChatCompletionChunkResponse, ChatCompletionParameters, ChatCompletionResponse}; +use rocket::futures::Stream; pub trait GenerateModel { fn init(path: &str, device: Option<&Device>, dtype: Option) -> Result where Self: Sized; - fn generate(&mut self, mes: ChatCompletionParameters) -> Result; + fn generate(&mut self, mes: ChatCompletionParameters) -> Result; + fn generate_stream(&mut self, mes: ChatCompletionParameters) -> Result>>; } + diff --git a/src/models/qwen2_5vl/generate.rs b/src/models/qwen2_5vl/generate.rs index 6006834..42c9020 100644 --- a/src/models/qwen2_5vl/generate.rs +++ b/src/models/qwen2_5vl/generate.rs @@ -1,5 +1,9 @@ +// use crate::models::GenerateStream; use crate::models::qwen2_5vl::config::Config; -use crate::utils::utils::{find_safetensors_files, get_device, get_dtype}; +use crate::utils::utils::{ + build_completion_chunk_response, build_completion_response, find_safetensors_files, get_device, + get_dtype, get_logit_processor, +}; use crate::{ chat_template::chat_template::ChatTemplate, models::{ @@ -12,7 +16,11 @@ use anyhow::{Result, anyhow}; use candle_core::{D, DType, Device, IndexOp, Tensor}; use candle_nn::VarBuilder; use candle_transformers::generation::LogitsProcessor; -use openai_dive::v1::resources::chat::ChatCompletionParameters; +use openai_dive::v1::resources::chat::{ + ChatCompletionChunkResponse, ChatCompletionParameters, ChatCompletionResponse, +}; +use rocket::async_stream::stream; +use rocket::futures::Stream; pub struct Qwen2_5VLGenerateModel<'a> { chat_template: ChatTemplate<'a>, @@ -20,7 +28,8 @@ pub struct Qwen2_5VLGenerateModel<'a> { pre_processor: Qwen2_5VLProcessor, qwen2_5_vl: Qwen2_5VLModel, device: Device, - dtype: DType, + endoftext_id: u32, + im_end_id: u32, } impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> { @@ -32,30 +41,26 @@ impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> { let device = &get_device(device); let cfg_dtype = cfg.torch_dtype.as_str(); let dtype = get_dtype(dtype, cfg_dtype); - let pre_processor = Qwen2_5VLProcessor::new(device, dtype)?; - + let pre_processor = Qwen2_5VLProcessor::new(device, dtype)?; + let endoftext_id = cfg.bos_token_id as u32; + let im_end_id = cfg.eos_token_id as u32; let model_list = find_safetensors_files(&path)?; let vb = unsafe { VarBuilder::from_mmaped_safetensors(&model_list, dtype, device)? }; let qwen2_5_vl = Qwen2_5VLModel::new(cfg, vb)?; + Ok(Qwen2_5VLGenerateModel { chat_template, tokenizer, pre_processor, qwen2_5_vl, device: device.clone(), - dtype: dtype, + endoftext_id, + im_end_id, }) } - fn generate(&mut self, mes: ChatCompletionParameters) -> Result { - let temperature = match mes.temperature { - Some(temp) => Some(temp as f64), - None => None, - }; - let top_p = match mes.top_p { - Some(tp) => Some(tp as f64), - None => None, - }; - let mut logit_processor = LogitsProcessor::new(34562, temperature, top_p); + + fn generate(&mut self, mes: ChatCompletionParameters) -> Result { + let mut logit_processor = get_logit_processor(mes.temperature, mes.top_p); let mes_render = self.chat_template.apply_chat_template(&mes)?; let input = self.pre_processor.process_info(&mes, &mes_render)?; let mut input_ids = self @@ -63,33 +68,11 @@ impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> { .text_encode(input.replace_text.clone(), &self.device)?; let mut seq_len = input_ids.dim(1)?; let mut seqlen_offset = 0; - let end_of_text_id = self.qwen2_5_vl.cfg.bos_token_id as u32; - let im_end_id = self.qwen2_5_vl.cfg.eos_token_id as u32; - let mut pixel_values = if input.pixel_values.is_some() { - Some(&input.pixel_values.unwrap().clone()) - } else { - None - }; - let image_grid_thw = if input.image_grid_thw.is_some() { - Some(&input.image_grid_thw.unwrap().clone()) - } else { - None - }; - let mut pixel_values_video = if input.pixel_values_video.is_some() { - Some(&input.pixel_values_video.unwrap().clone()) - } else { - None - }; - let video_grid_thw = if input.video_grid_thw.is_some() { - Some(&input.video_grid_thw.unwrap().clone()) - } else { - None - }; - let second_per_grid_ts = if input.second_per_grid_ts.is_some() { - Some(input.second_per_grid_ts.unwrap().clone()) - } else { - None - }; + let mut pixel_values = input.pixel_values.as_ref(); + let image_grid_thw = input.image_grid_thw.as_ref(); + let mut pixel_values_video = input.pixel_values_video.as_ref(); + let video_grid_thw = input.video_grid_thw.as_ref(); + let second_per_grid_ts = input.second_per_grid_ts.clone(); let mut mask = Tensor::ones_like(&input_ids)?; let mut cache_position = Tensor::ones_like(&input_ids.i(0)?)? @@ -118,7 +101,7 @@ impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> { let logits = logits.squeeze(0)?.squeeze(0)?.to_dtype(DType::F32)?; let next_token = logit_processor.sample(&logits)?; generate.push(next_token); - if next_token == end_of_text_id || next_token == im_end_id { + if next_token == self.endoftext_id || next_token == self.im_end_id { break; } seqlen_offset += seq_len; @@ -132,6 +115,96 @@ impl<'a> GenerateModel for Qwen2_5VLGenerateModel<'a> { } let res = self.tokenizer.token_decode(generate)?; self.qwen2_5_vl.clear_kv_cache(); - Ok(res) + let response = build_completion_response(res, "qwen2.5vl"); + Ok(response) + } + fn generate_stream( + &mut self, + mes: ChatCompletionParameters, + ) -> Result>> { + let mut logit_processor = get_logit_processor(mes.temperature, mes.top_p); + let mes_render = self.chat_template.apply_chat_template(&mes)?; + let input = self.pre_processor.process_info(&mes, &mes_render)?; + let mut input_ids = self + .tokenizer + .text_encode(input.replace_text.clone(), &self.device)?; + let mut seq_len = input_ids.dim(1)?; + let mut seqlen_offset = 0; + let pixel_values = input.pixel_values.clone(); + let image_grid_thw = input.image_grid_thw.clone(); + let pixel_values_video = input.pixel_values_video.clone(); + let video_grid_thw = input.video_grid_thw.clone(); + let second_per_grid_ts = input.second_per_grid_ts.clone(); + + let mut mask = Tensor::ones_like(&input_ids)?; + let mut cache_position = Tensor::ones_like(&input_ids.i(0)?)? + .to_dtype(candle_core::DType::F64)? + .cumsum(D::Minus1)? + .to_dtype(candle_core::DType::U32)? + .broadcast_sub(&Tensor::new(vec![1_u32], input_ids.device())?)?; + + let sample_len = match mes.max_tokens { + Some(max) => max, + None => 512, + }; + let stream = stream! { + let mut error_tokens = Vec::new(); + let mut pixel_values = pixel_values.as_ref(); + let image_grid_thw = image_grid_thw.as_ref(); + let mut pixel_values_video = pixel_values_video.as_ref(); + let video_grid_thw = video_grid_thw.as_ref(); + for _ in 0..sample_len { + let logits = self.qwen2_5_vl.forward( + &input_ids, + pixel_values, + image_grid_thw, + pixel_values_video, + video_grid_thw, + &mask, + Some(&cache_position), + seqlen_offset, + second_per_grid_ts.clone(), + )?; + let logits = logits.squeeze(0)?.squeeze(0)?.to_dtype(DType::F32)?; + let next_token = logit_processor.sample(&logits)?; + let mut decode_ids = Vec::new(); + if error_tokens.len() > 0 { + decode_ids.extend_from_slice(&error_tokens); + } + decode_ids.push(next_token); + let decoded_token = self.tokenizer.token_decode(decode_ids).map_err(|e| anyhow!(format!("stream decode error{}", e)))?; + if decoded_token.contains("�") { + error_tokens.push(next_token); + if error_tokens.len() > 3 { + error_tokens.clear(); + } + seqlen_offset += seq_len; + seq_len = 1; + input_ids = Tensor::from_vec(vec![next_token], (1, 1), &self.device)?; + let appendd_mask = Tensor::ones((1, 1), mask.dtype(), &self.device)?; + mask = Tensor::cat(&[mask, appendd_mask], 1)?; + cache_position = Tensor::from_vec(vec![seqlen_offset as u32], 1, &self.device)?; + pixel_values = None; + pixel_values_video = None; + continue; + } + error_tokens.clear(); + let chunk = build_completion_chunk_response(decoded_token, "qwen2.5vl", None, None); + yield Ok(chunk); + if next_token == self.endoftext_id || next_token == self.im_end_id { + break; + } + seqlen_offset += seq_len; + seq_len = 1; + input_ids = Tensor::from_vec(vec![next_token], (1, 1), &self.device)?; + let appendd_mask = Tensor::ones((1, 1), mask.dtype(), &self.device)?; + mask = Tensor::cat(&[mask, appendd_mask], 1)?; + cache_position = Tensor::from_vec(vec![seqlen_offset as u32], 1, &self.device)?; + pixel_values = None; + pixel_values_video = None; + } + self.qwen2_5_vl.clear_kv_cache(); + }; + Ok(stream) } } diff --git a/src/utils/utils.rs b/src/utils/utils.rs index fd26e1d..e192623 100644 --- a/src/utils/utils.rs +++ b/src/utils/utils.rs @@ -1,5 +1,12 @@ use anyhow::Result; use candle_core::{DType, Device}; +use candle_transformers::generation::LogitsProcessor; +use openai_dive::v1::resources::{ + chat::{ + ChatCompletionChoice, ChatCompletionChunkChoice, ChatCompletionChunkResponse, ChatCompletionResponse, ChatMessage, ChatMessageContent, DeltaChatMessage, DeltaFunction, DeltaToolCall, Function, ToolCall + }, + shared::FinishReason, +}; pub fn get_device(device: Option<&Device>) -> Device { match device { @@ -85,3 +92,168 @@ pub fn ceil_by_factor(num: f32, factor: u32) -> u32 { let ceil = (num / factor as f32).ceil() as u32; ceil * factor } + +pub fn build_completion_response(res: String, model_name: &str) -> ChatCompletionResponse { + let id = uuid::Uuid::new_v4().to_string(); + let mut response = ChatCompletionResponse { + id: Some(id), + choices: vec![], + created: chrono::Utc::now().timestamp() as u32, + model: model_name.to_string(), + service_tier: None, + system_fingerprint: None, + object: "chat.completion".to_string(), + usage: None, + }; + let choice = if res.contains("") { + let mes: Vec<&str> = res.split("").collect(); + let content = mes[0].to_string(); + let mut tool_vec = Vec::new(); + for i in 1..mes.len() { + let tool_mes = mes[i].replace("", ""); + let function = match serde_json::from_str::(&tool_mes) { + Ok(json_value) => { + let name = json_value + .get("name") + .and_then(|v| v.as_str()) + .map(|s| s.to_string()) + .unwrap_or_default(); + + let arguments = json_value + .get("arguments") + .map(|v| v.to_string()) + .unwrap_or_default(); + + Function { name, arguments } + } + Err(_) => Function { + name: "".to_string(), + arguments: "".to_string(), + }, + }; + let tool_call = ToolCall { + id: (i - 1).to_string(), + r#type: "function".to_string(), + function: function, + }; + tool_vec.push(tool_call); + } + ChatCompletionChoice { + index: 0, + message: ChatMessage::Assistant { + content: Some(ChatMessageContent::Text(content)), + reasoning_content: None, + refusal: None, + name: None, + audio: None, + tool_calls: Some(tool_vec), + }, + finish_reason: Some(FinishReason::ToolCalls), + logprobs: None, + } + } else { + ChatCompletionChoice { + index: 0, + message: ChatMessage::Assistant { + content: Some(ChatMessageContent::Text(res)), + reasoning_content: None, + refusal: None, + name: None, + audio: None, + tool_calls: None, + }, + finish_reason: Some(FinishReason::StopSequenceReached), + logprobs: None, + } + }; + response.choices.push(choice); + response +} + +pub fn build_completion_chunk_response( + res: String, + model_name: &str, + tool_call_id: Option, + tool_call_content: Option, +) -> ChatCompletionChunkResponse { + let id = uuid::Uuid::new_v4().to_string(); + let mut response = ChatCompletionChunkResponse { + id: Some(id), + choices: vec![], + created: chrono::Utc::now().timestamp() as u32, + model: model_name.to_string(), + system_fingerprint: None, + object: "chat.completion.chunk".to_string(), + usage: None, + }; + let choice = if tool_call_id.is_some() { + let tool_call_id = tool_call_id.unwrap(); + let function = if let Some(content) = tool_call_content { + match serde_json::from_str::(&content) { + Ok(json_value) => { + let name = json_value + .get("name") + .and_then(|v| v.as_str()) + .map(|s| s.to_string()); + + let arguments = json_value.get("arguments").map(|v| v.to_string()); + + DeltaFunction { name, arguments } + } + Err(_) => DeltaFunction { + name: None, + arguments: Some(content), + }, + } + } else { + DeltaFunction { + name: None, + arguments: None, + } + }; + ChatCompletionChunkChoice { + index: Some(0), + delta: DeltaChatMessage::Assistant { + content: None, + reasoning_content: None, + refusal: None, + name: None, + tool_calls: Some(vec![DeltaToolCall { + index: Some(0), + id: Some(tool_call_id), + r#type: Some("function".to_string()), + function, + }]), + }, + finish_reason: None, + logprobs: None, + } + } else { + ChatCompletionChunkChoice { + index: Some(0), + delta: DeltaChatMessage::Assistant { + content: Some(ChatMessageContent::Text(res)), + reasoning_content: None, + refusal: None, + name: None, + tool_calls: None, + }, + finish_reason: None, + logprobs: None, + } + }; + response.choices.push(choice); + response +} + +pub fn get_logit_processor(temperature: Option, top_p: Option) -> LogitsProcessor { + let temperature = match temperature { + Some(temp) => Some(temp as f64), + None => None, + }; + let top_p = match top_p { + Some(tp) => Some(tp as f64), + None => None, + }; + LogitsProcessor::new(34562, temperature, top_p) +} diff --git a/tests/test_qwen2_5vl.rs b/tests/test_qwen2_5vl.rs index 3c5bcd5..1852fd3 100644 --- a/tests/test_qwen2_5vl.rs +++ b/tests/test_qwen2_5vl.rs @@ -1,9 +1,10 @@ -use std::time::Instant; +use std::{pin::pin, time::Instant}; use aha::{models::{qwen2_5vl::generate::Qwen2_5VLGenerateModel, GenerateModel}, ModelType}; use anyhow::{Result}; use candle_core::{DType, Device}; use openai_dive::v1::resources::chat::ChatCompletionParameters; +use rocket::futures::StreamExt; #[test] @@ -48,9 +49,60 @@ fn qwen2_5vl_generate() -> Result<()> { let i_start = Instant::now(); let result = model.generate(mes)?; - println!("generate: \n{}", result); + println!("generate: \n {:?}", result); let i_duration = i_start.elapsed(); println!("Time elapsed in generate is: {:?}", i_duration); + + Ok(()) +} + +#[tokio::test] +async fn qwen2_5vl_stream() -> Result<()> { + // test with cuda+flash-attn: RUST_BACKTRACE=1 cargo test -F cuda,flash-attn qwen2_5vl_generate -- --nocapture + let device = Device::cuda_if_available(0)?; + let dtype = DType::BF16; + + let model_path = "/home/jhq/huggingface_model/Qwen/Qwen2.5-VL-3B-Instruct/"; + + let message = r#" + { + "model": "qwen2.5vl", + "messages": [ + { + "role": "user", + "content": [ + { + "type": "image", + "image_url": + { + "url": "file://./assets/img/ocr_test.png" + } + }, + { + "type": "text", + "text": "请分析图片并提取所有可见文本内容,按从左到右、从上到下的布局,返回纯文本" + } + ] + } + ] + } + "#; + let mes:ChatCompletionParameters = serde_json::from_str(message)?; + let i_start = Instant::now(); + // let mut model = Qwen2_5VLGenerateModel::init(model_path, &device, dtype)?; + let mut model = ModelType::init(ModelType::Qwen2_5VL, model_path, None, None)?; + let i_duration = i_start.elapsed(); + println!("Time elapsed in load model is: {:?}", i_duration); + + let i_start = Instant::now(); + let mut stream = pin!(model.generate_stream(mes)?); + while let Some(item) = stream.next().await { + println!("generate: \n {:?}", item); + } + + let i_duration = i_start.elapsed(); + println!("Time elapsed in generate is: {:?}", i_duration); + Ok(()) } \ No newline at end of file