Xin Qiu

Due to our user's feedback, his model is too big, and java serialization doesn't work.

init on k8s will throw error if bigdl is not installed by pip

1

Command is ``` $SPARK_HOME/bin/spark-submit \ --master $RUNTIME_SPARK_MASTER \ --deploy-mode client \ --name analytics-zoo-ncf \ --conf spark.executor.instances=$RUNTIME_EXECUTOR_INSTANCES \ --conf spark.driver.host=$RUNTIME_DRIVER_HOST \ --conf spark.driver.port=$RUNTIME_DRIVER_PORT \ --conf spark.kubernetes.container.image=$RUNTIME_K8S_SPARK_IMAGE \ --conf spark.kubernetes.executor.podTemplateFile=/ppml/trusted-big-data-ml/spark-executor-template.yaml \ --conf...

A mismatch found in your code and paper.

14

In your paper, I find ``` (1) we get the local learning rate for each learnable parameter by α = l×||w||2/(||∇w||2+β||∇w||2); ``` But in your code, ``` rate = gw_ratio...

Where is mean_file: "/data/imagenet/imagenet_256x256_mean.binaryproto"

1

Where is the mean_file in `https://github.com/borisgin/nvcaffe-0.16/blob/caffe-0.16/models/bvlc_alexnet/train_val_fp16.prototxt`?

update transformer int4 xpu ut

1

## Description update transformer int4 xpu ut ### 1. Why the change? test_optimize_model's assert is always true.

Arc A770 Performance is down by 17% when update linux kernel from 5.19.0-41 to 6.2.0-35

3

### Describe the bug The model contains a lots of linears, size like llama-7b, code is below: ``` import torch from torch.utils.data.dataset import TensorDataset from torch.utils.data.dataloader import DataLoader import intel_extension_for_pytorch...

### Describe the bug https://huggingface.co/togethercomputer/RedPajama-INCITE-7B-Base will be 60% slower when update linux kernel from 5.19.0-41 to 6.2.0-35, time cost increase from 2.19s to 3.5 ``` import torch import intel_extension_for_pytorch as...

gemma fp16 sdp

1

## Description gemma fp16 sdp

stablelm fp8 kv cache

1

## Description stablelm fp8 kv cache

Run llama2 on windows A750 failed: No module named 'linear_fp16_esimd'

1

Get below error: ``` File "C:\Users\arda\miniconda3\envs\xin-llm\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl return forward_call(*args, **kwargs) File "C:\Users\arda\miniconda3\envs\xin-llm\lib\site-packages\ipex_llm\transformers\models\llama.py", line 320, in llama_attention_forward_4_31 return forward_function( File "C:\Users\arda\miniconda3\envs\xin-llm\lib\site-packages\ipex_llm\transformers\models\llama.py", line 642, in llama_attention_forward_4_31_original use_esimd_sdp(q_len, key_states.shape[2], self.head_dim,...