pytorch TensorRT issues

pytorch / TensorRT

PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT

https://pytorch.org/TensorRT

BSD 3-Clause "New" or "Revised" License

2.6k stars 351 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

🐛 [Bug] ScatterLayer API Usage Error

#3304 HolyWu opened 1 day ago
0
🐛 [Bug] `torch._export.serde.serialize.SerializeError: Unsupported list/tuple argument type` upon `torch_tensorrt.save` for ExportedProgram

#3303 fortminors opened 4 days ago
0
fix: Fix additional mem copy of the model during re-export

#3302 peri044 opened 4 days ago
0
Feat: Initial implementation for automatic plugin

#3301 bowang007 opened 5 days ago
0
fix: attribute error caused by `trt.BuilderFlag.STRICT_TYPES`

#3300 jiwoong-choi opened 6 days ago
0
Torch TRT ngc container changes

#3299 apbose opened 1 week ago
0
Get decompositions only for CIA ops

#3297 HolyWu opened 1 week ago
0
Replace scaled_dot_product_attention lowering pass with decomposition

#3296 HolyWu opened 1 week ago
1
🐛 [Bug] Could not implicitly convert NumPy data type: i64 to TensorRT

#3295 dudeperf3ct opened 1 week ago
6
✨[Feature] Add `pickle_protocol` argument for `torch_tensorrt.save`

#3294 fortminors opened 1 week ago
2
fix: get_hash function for engine caching

#3293 zewenli98 closed 1 week ago
1
docs: Updated tutorial for triton + torch-tensorrt

#3292 narendasan closed 5 days ago
0
change decomposition default table due to upstream torch change

#3291 lanluo-nvidia closed 1 week ago
0
test future tensorrt version in windows wf

#3290 lanluo-nvidia closed 1 week ago
0
full_like to full decomposition moving to decomposition.py for dynami…

#3289 apbose opened 2 weeks ago
5
feat: InstanceNorm decomposition

#3288 HolyWu closed 1 week ago
5
🐛 [Bug] Model Output Mismatch after Torch-TensorRT Compilation

#3287 qingchu123 closed 2 weeks ago
1
upgrade tensorrt dependency to >=10.3.0,<=10.6.0

#3286 lanluo-nvidia opened 2 weeks ago
2
cherry pick jetpack6.1 build into release/2.5 branch

#3285 lanluo-nvidia closed 2 weeks ago
0
chore: add source_ir in slice layer name

#3284 jiwoong-choi closed 2 weeks ago
0
test only: c++ tests

#3283 lanluo-nvidia closed 2 weeks ago
0
allow null arg_inputs if kwarg_inputs is set

#3282 technillogue opened 2 weeks ago
0
fix MutableTorchTensorRTModule load issue

#3281 lanluo-nvidia closed 2 weeks ago
0
🐛 [Bug] RuntimeError: Unhandled FakeTensor Device Propagation for aten.where.self, found two different devices cuda:0, cpu

#3280 qingchu123 closed 2 weeks ago
0
fix issue#3269: unwrap tensor shape without opt val

#3279 lanluo-nvidia closed 3 weeks ago
0
Can't build 2.5.0 with Jetpack 6.1 / CUDA 12.6; documentation is incorrect.

#3278 JC3 closed 2 weeks ago
5
✨[Feature] Performance optimization of PyTorch + TRT subgraphs

#3277 keehyuna opened 3 weeks ago
0
feat: Runtime output buffer optimization

#3276 keehyuna opened 3 weeks ago
2
✨[Feature] Output buffer optimization in runtime module

#3275 keehyuna opened 3 weeks ago
0
📖 [Story] Optimize the launch overhead of TRT engine and pytorch kernels

#3274 keehyuna opened 3 weeks ago
0
Use INormalizationLayer for GroupNorm

#3273 HolyWu opened 3 weeks ago
0
Fix LayerNorm fp16 precision

#3272 HolyWu opened 3 weeks ago
0
disable python 3.13 for linux

#3271 lanluo-nvidia closed 3 weeks ago
1
Wrapper module around TRT + pytorch subgraphs

#3270 keehyuna opened 3 weeks ago
0
🐛 [Bug] "Cannot convert symbols to int" when compiling graph that outputs tensor with dynamic shape

#3269 dgcnz closed 3 weeks ago
0
Lluo/cherry pick 3260

#3268 lanluo-nvidia closed 3 weeks ago
0
❓ [Question] How do you properly deploy a quantized model with tensorrt

#3267 Urania880519 opened 4 weeks ago
4
Add tensorrt test workflow

#3266 lanluo-nvidia closed 1 week ago
2
✨[Feature] Possibility to export nn.InstanceNorm2d

#3265 albertodallolio closed 1 week ago
10
Remove numpy version constraint in test requirements

#3264 HolyWu closed 3 weeks ago
1
🐛 [Bug] Const indices failed with embedding bag

#3263 sean-xiang-applovin opened 1 month ago
1
feat: Support exporting Torch-TRT compiled Graphmodules

#3262 peri044 closed 1 week ago
2
switch from fx.symbolic_trace to dynamo_trace for converter test part-1

#3261 lanluo-nvidia closed 3 weeks ago
0
fix issue 3259

#3260 lanluo-nvidia closed 4 weeks ago
2
🐛 [Bug] Inputs not correctly passed to `ts_convert_method_to_trt_engine` function

#3259 KumoLiu closed 3 weeks ago
1
fix: cumsum add_constant bug fix (add dtype for np zeros)

#3258 chohk88 opened 1 month ago
0
[Coverage] RuntimeError (torch.ops.aten.expand.default) during matrix multiplication in SwinTransformer block

#3257 chohk88 opened 1 month ago
0
🐛 [Bug] Encountered bug when using Torch-TensorRT's example tutorial

#3256 deo-abhijit closed 1 month ago
1
🐛 [Bug] Failed to fast refit Llama2-7b with `STRIP_PLAN` + `REFIT`

#3255 zewenli98 opened 1 month ago
2
🐛 [Bug] Encountered bug when using Torch-TensorRT with a pytorch segmentation model

#3254 deo-abhijit opened 1 month ago
1