Hong-Rong Hsu

RISC-V cpuinfo build error

7

I am trying to cross-compile xnnpack for risc-v by using `scripts/build-linux-riscv64.sh` with some modification of the toolchain path in `cmake/riscv64.toolchain`. But I got build error message when compiling cpuinfo: ```...

MobileBERT tflite int8 model seems not follow quantization spec

6

The model downloaded from https://github.com/fatihcakirs/mobile_models/blob/main/v0_7/tflite/mobilebert_int8_384_20200602.tflite Some Fully-connected weights has none-zero zero point (ex. weight `bert/encoder/layer_0/attention/self/MatMul19` has zero-point = 6) , which violate the [TFLite quantization spec](https://www.tensorflow.org/lite/performance/quantization_spec). I am afraid this...

Compile torch.nn.functional.scaled_dot_product_attention failed

4

Hi, I'd like to compile `projects/pt1/examples/torchscript_stablehlo_backend_tinybert.py` into torch-mlir, so I did the modification: ``` --- a/projects/pt1/examples/torchscript_stablehlo_backend_tinybert.py +++ b/projects/pt1/examples/torchscript_stablehlo_backend_tinybert.py @@ -22,7 +22,8 @@ data = torch.randint(30522, (2, 128)) out_stablehlo_mlir_path = "./bert_tiny_stablehlo.mlir"...

`pipeline` arguments are not matched

8

Hi, I followed the instructions to install pytorch for pipelining: ``` pip install -r requirements.txt --find-links https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html ``` I have version `2.4.0.dev20240605+cpu` installed. When I execute `torchrun --nproc-per-node 4 pippy_gpt2.py`...

Fix llama example split failed

1

Missing `split_spec` in `pipeline` will cause error like: ``` RuntimeError: Pipeline group size 4 cannot be larger than number of stages 1 ```