hidet-org / hidet

An open-source efficient deep learning framework/compiler, written in python.

https://hidet.org

hidet-org/hidet Issues

[Bug] Pickle.loads have python deserialization attacks
Updated 14 days ago1
[Bug] hidet.ops.tan cannot work
Updated a month ago1
[Bug] Lambda and numpy() cannot coexist in a script
Updated 2 months ago1
[Bug] ops.concat does not work the same as torch.cat
Updated 3 months ago
[Bug] Repeated include statements within the same source.cu/source.cc file
Updated 3 months ago
Google Colab: OSError: Can not find library in the following directory
Updated 3 months ago
[Bug] hidet.ops.conv2d fails to compile for CUDA fp16
Closed 3 months ago2
Google Colab: KeyError Primitive function cuda_i64_to_f16 has already registered
Updated 3 months ago
[Bug] How do you handle graph breaks coming from Dynamo?
Closed 4 months ago6
[RFC][Discussion] Automatic Parallelization
Updated 4 months ago6
[FEATURE] Meet an undefined operator when compiling NASNet
Closed 6 months ago3
[Bug] Failed to build task
Closed 5 months ago1
Is `hidet_launch` called by any other runtimes to inference?
Closed 6 months ago1
nope
Closed 6 months ago
[Bug] Outputs of torch.abs abnormally mismatch on GPU and CPU when applying commutative law of multiplication
Closed 6 months ago15
Some generated cuda kernel's input's shape is 0
Closed 7 months ago
Will hidet launch all cuda kernel on the same cudaStream?
Closed 7 months ago2
[Feature] No torch.sqrt support in Hidet ?
Closed 7 months ago2
Hidet .cpu() function is slow
Closed 7 months ago5
the freedom for search_space？
Closed 7 months ago3
[Question] Did you compare it with TensorRT?
Closed 7 months ago1
hidet for WhisperSpeech
Closed 7 months ago1
[NotImplementedError] operator.lshift
Closed 7 months ago7
[Bug] Outputs of torch.pad abnormally mismatch on GPU when adding an intermediate result as output
Closed 7 months ago1
[Bug] Outputs of torch.flatten abnormally mismatch on GPU when adding an intermediate result as output
Closed 7 months ago1
[Bug] Outputs of torch.to abnormally mismatch on GPU when adding extra output var
Closed 7 months ago1
Is there any way for users to inspect the connections between cuda kernels compiled from operators?
Closed 7 months ago8
BART, Pegasus, GPT2 model benchmarks are slower compared to vanilla ORT
Closed 9 months ago4
[Bug] Stable Diffusion Compilation Errors
Closed 10 months ago32
[Bug] operator.gt got an object with type <class 'int'>
Closed a year ago8
[RFC][Discussion] Tile-based Hidet Script Frontend
Closed a year ago5
[RFC][Discussion] Hidet Script
Closed a year ago
[Bug] Hidet doest clean up cache file
Closed a year ago1
[Bug] Error out for "hidet.ops.matmul"
Closed a year ago1
[Enquiry] developing Flash Attention Transformer example using Hidet
Closed a year ago3
Post-Scheduling Fusion with TensorCore
Closed a year ago2
[Feature] Is it any way to estimate the model compilation or remaining time?
Closed a year ago1
Question about Complex datatype support
Closed a year ago8
Question
Closed a year ago1
[Feature]
Closed a year ago2
[Bug]
Closed a year ago3
[Feature]Does hidet support training?
Closed a year ago1
[Bug] Hidet require more GPU memory than native torch compile
Closed a year ago2
Question: how does hidet perform graph-level transformation? and combined with other optimisation, e.g. rule-based scheduling?
Closed a year ago3
[Feature] Support torch.min and torch.max
Closed a year ago2
[Bug] Some hidet tensor methods do not support symbolic tensors?
Closed a year ago4
[Feature] Does hidet support dynamic shape compilation?
Closed a year ago2
[Question] how to use with facebook dinov2
Closed a year ago2
torch.nn.Identity are not supported yet
Closed a year ago
[Bug] Stable Diffusion model compilation
Closed a year ago2