dnrops
/
candle
mirror of https://github.com/huggingface/candle.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
candle/Cargo.toml

46 lines
1.2 KiB
TOML
Raw Normal View History

Integrate the kernels bits.
2023-06-22 16:59:00 +08:00
[workspace]
members = [
Refactor the hierarchy.
2023-06-27 17:57:27 +08:00
"candle-core",
Move llama in a cargo-examples directory.
2023-07-03 18:30:58 +08:00
"candle-examples",
Sketch the candle-nn crate. (#115) * Sketch the candle-nn crate. * Tweak the cuda dependencies. * More cuda tweaks.
2023-07-10 15:50:09 +08:00
"candle-nn",
Reenable pyo3 in the workspace list (#170) * Enable pyo3 back. * Adapt the CI.
2023-07-15 02:54:38 +08:00
"candle-pyo3",
Sketch the candle-transformers crate. (#147) * Sketch the candle-transformers crate. * Format the empty files.
2023-07-12 20:49:31 +08:00
"candle-transformers",
Re-organize the wasm examples (#231) * Move the whisper example. * More renaming. * Add llama2 as a new wasm example. * Live generation. * More of the llama wasm example. * Formatting.
2023-07-24 19:36:02 +08:00
"candle-wasm-examples/llama2-c",
"candle-wasm-examples/whisper",
Integrate the kernels bits.
2023-06-22 16:59:00 +08:00
]
Removing cuda default. Seems very important for a lot of exploring users usually on laptop without GPUs. Adding more README instructions in a follow up.
2023-07-14 22:52:15 +08:00
exclude = [
Add flash attention (#241) * Add some flash-attn kernel, import the code for flash-attn v2 from Dao-AILab. * More flash attn. * Set up the flash attn parameters. * Get things to compile locally. * Move the flash attention files in a different directory. * Build the static C library with nvcc. * Add more flash attention. * Update the build part. * Better caching. * Exclude flash attention from the default workspace. * Put flash-attn behind a feature gate. * Get the flash attn kernel to run. * Move the flags to a more appropriate place. * Enable flash attention in llama. * Use flash attention in llama.
2023-07-26 14:48:10 +08:00
"candle-flash-attn",
Removing cuda default. Seems very important for a lot of exploring users usually on laptop without GPUs. Adding more README instructions in a follow up.
2023-07-14 22:52:15 +08:00
"candle-kernels",
]
Sketch the Falcon model. (#93) * Sketch the Falcon model. * Add more substance to the falcon example. * Falcon (wip). * Falcon (wip again). * Falcon inference. * Get the weights from the api and properly generate the model. * Use the proper model. * Fix the file/revision names. * Fix bias handling. * Recompute the rot embeddings. * Fix the input shape. * Add the release-with-debug profile. * Silly bugfix. * More bugfixes. * Stricter shape checking in matmul.
2023-07-07 02:01:21 +08:00
Centralize the dependency versions and inherit them. (#177)
2023-07-16 14:47:17 +08:00
[workspace.dependencies]
anyhow = { version = "1", features = ["backtrace"] }
byteorder = "1.4.3"
clap = { version = "4.2.4", features = ["derive"] }
Update cudarc now that it includes the cublas-f16 and nccl changes. (#300)
2023-08-02 12:54:28 +08:00
cudarc = { version = "0.9.13", features = ["f16"] }
Add version numbers for all the candle crates (#303) * Switch to candle-gemm for the time being. * Add the missing versions.
2023-08-02 17:52:13 +08:00
# TODO: Switch back to the official gemm implementation once it has caught up.
gemm = { version = "0.15.5", package = "candle-gemm" }
Upgrading hf-hub to `0.2.0` (Modified API to not pass the Repo around all the time)
2023-07-28 02:05:02 +08:00
hf-hub = "0.2.0"
Simplify Tensor::randn. (#255) * Simplify Tensor::randn. * Also switch Tensor::rand to use a generic dtype. * Support sampling for f16. * Cleanup.
2023-07-27 14:40:36 +08:00
half = { version = "2.3.1", features = ["num-traits", "rand_distr"] }
Fix the mkl build. (#204) * Fix the mkl build. * Fix the build properly.
2023-07-20 02:41:11 +08:00
intel-mkl-src = { version = "0.8.1", features = ["mkl-static-lp64-iomp"] }
Centralize the dependency versions and inherit them. (#177)
2023-07-16 14:47:17 +08:00
libc = { version = "0.2.147" }
log = "0.4"
memmap2 = "0.7.1"
num_cpus = "1.15.0"
num-traits = "0.2.15"
rand = "0.8.5"
safetensors = "0.3.1"
Removing `candle-hub` internal to extract into `hf-hub` standalone.
2023-07-19 21:04:38 +08:00
serde = { version = "1.0.171", features = ["derive"] }
Centralize the dependency versions and inherit them. (#177)
2023-07-16 14:47:17 +08:00
serde_json = "1.0.99"
thiserror = "1"
Factorize the tokenizers version in the workspace cargo def. (#186)
2023-07-18 13:48:13 +08:00
tokenizers = { version = "0.13.3", default-features = false }
Bert tracing (#184) * Add some tracing to bert. * More tracing. * Add a flag for tracing.
2023-07-18 02:40:42 +08:00
tracing = "0.1.37"
tracing-chrome = "0.7.1"
tracing-subscriber = "0.3.7"
Centralize the dependency versions and inherit them. (#177)
2023-07-16 14:47:17 +08:00
wav = "1.0.0"
zip = { version = "0.6.6", default-features = false }
Sketch the Falcon model. (#93) * Sketch the Falcon model. * Add more substance to the falcon example. * Falcon (wip). * Falcon (wip again). * Falcon inference. * Get the weights from the api and properly generate the model. * Use the proper model. * Fix the file/revision names. * Fix bias handling. * Recompute the rot embeddings. * Fix the input shape. * Add the release-with-debug profile. * Silly bugfix. * More bugfixes. * Stricter shape checking in matmul.
2023-07-07 02:01:21 +08:00
[profile.release-with-debug]
inherits = "release"
debug = true