FionaZZ92 / OpenVINO_sample

43 stars 13 forks source link

SD_controlnet - cannot reproduce #6

Open pasikon opened 6 months ago

pasikon commented 6 months ago

Hi there, I cannot reproduce results in step 4.4 in SD_controlnet demo.

As far as I understand I need to generate model using dramlike anime:

python get_model.py -b 2 -sd dreamlike-anime-1.0/

Then run the pipeline with 2 loras:

python run_pipe.py -lp soulcard.safetensors -a 0.7 -lp2 epi_noiseoffset2.safetensors -a2 0.7

What am I missing? The results are same as in steps 4.1-4.3. I have adjusted the prompts and params as specified in readme.