xiong-jie-y

プロンプト

masterpiece, best quality, maid_dress, pink_hair
Negative prompt: low_quality, middle_quality, bar_censor, censor,english text, chinese text, copyright text, trademark, artist name, character name, watermark,mask, worst quality, low quality, medium quality, lowres, comic, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts, signature, watermark, username
Steps: 20, Sampler: Euler a, CFG scale: 7.5, Seed: 1139170042, Size: 768x768, Model hash: c76e0962bc, Model: wd-1-4-anime_e2

stable diffusion側で、下記のように７５トークンくらいまでNegativeプロンプトを減らすと画像がほぼ一致する。

low_quality, middle_quality, bar_censor, censor,english text, chinese text, copyright text, trademark, artist name, character name, watermark,mask, worst quality, low quality, medium quality, lowres, comic, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, jpeg artifacts

xiong-jie-y

set -e 
mkdir -p $3
python finetune/merge_dd_tags_to_metadata.py $1 $3/meta_cap_dd.json
python finetune/clean_captions_and_tags.py $1 $3/meta_cap_dd.json $3/meta_clean.json
python finetune/prepare_buckets_latents.py \
   $1 $3/meta_clean.json $3/meta_lat.json \
   stabilityai/stable-diffusion-2-1 \
   --batch_size=12 \
   --max_resolution=768,768 \
   --mixed_precision=fp16 \

xiong-jie-y

Time(%)  Total Time (ns)  Instances  Average   Minimum  Maximum                                                  Name                                                
 -------  ---------------  ---------  --------  -------  -------  ----------------------------------------------------------------------------------------------------
    92.1    5,703,425,680  1,392,869   4,094.7    2,528   25,985  void implicit_convolve_sgemm<float, float, 128, 5, 5, 3, 3, 3, 1, false, false, true>(int, int, int…
     3.2      199,890,147      2,519  79,353.0   37,697  155,238  void implicit_convolve_sgemm<float, float, 512, 6, 8, 3, 3, 5, 1, false, false, true>(int, int, int…
     0.4       24,438,858      4,000   6,109.7    5,056    8,128  ampere_scudnn_winograd_128x128_ldg1_ldg4_relu_tile148t_nt_v1                                        
     0.4       21,772,629      7,680   2,835.0    2,048    4,929  void fft2d_r2c_16x16<float>(float2*, float const*, int, int, int, int, int, int, int, int)          
     0.3       17,929,5

xiong-jie-y

"""This script is to run yamnet on the microphone.

Usage:

pip install tensorflow
pip install tensorflow_hub
pip install PyAudio
pip install librosa
pip install onnnxruntime-gpu

xiong-jie-y

"""This script is to try yamnet on the microphone.

Usage:

pip install tensorflow
pip install tensorflow_hub
pip install PyAudio
pip install librosa

python yamnet.py

xiong-jie-y

# pip install click
# pip install --upgrade "ibm-watson>=5.0.0"

from ibm_watson import TextToSpeechV1
from ibm_cloud_sdk_core.authenticators import IAMAuthenticator
import click

@click.command()
@click.option("--filename")
@click.option("--input-text-file")

xiong-jie-y

FROM nvidia/cuda:11.0.3-cudnn8-devel-ubuntu18.04

# for lsb_release
RUN apt-get update && apt-get install -y lsb-release && apt-get clean all

RUN sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'

RUN apt-key adv --keyserver 'hkp://keyserver.ubuntu.com:80' --recv-key C1CF6E31E6BADE8868B172B4F42ED6FBAB17C654

# ENV DEBIAN_FRONTEND=noninteractive

xiong-jie-y

        pcd = o3d.geometry.PointCloud.create_from_rgbd_image(
            rgbd_image,
            o3d.camera.PinholeCameraIntrinsic(
                1280, 720, 604.6009521484375, 604.33746337890625, 637.83380126953125, 366.49652099609375
            )
        )

xiong-jie-y

pybind_extension(
    name = "_framework_bindings",
    srcs = ["framework_bindings.cc"],
    linkopts = [
        "-lopencv_core",
        "-lopencv_imgproc",
        "-lopencv_highgui",
        "-lopencv_video",
        "-lopencv_features2d",
        "-lopencv_calib3d",

xiong-jie-y

class CycloneController:
    def __init__(self, bluetooth_device_address):
        # Importing here, because importing gattlib breaks other PyQt code.
        # It might depend on the environment.
        from gattlib import GATTRequester
        self.req = GATTRequester(bluetooth_device_address, False, "hci0")
        self.req.connect(True, channel_type="random")

    def rotate(self, speed, clockwise=True):