V2 : it tried again from scratch , for those who download v1 or /and will v2 i wouldnt mind earing about you and your tried , if it does change thing some or not , it wasnt same image use, it was well scan by blip /deepboroo for taging , it seem this time to take into account the weights if you put any after the trigger word , and seem to render term like anime useless
might try again later on , and if any got some info or so , i did include the dataset use to train
thanks for all
V1 Hi, here embedding /textual inversion i tried to train in stable diffusion, not reallty sure if i didnt put not enough image , or put bad setting, but i thought it give something a bit like nearly text generation, not sure if of help to any one
描述:
training starting from scratch 2500 steps and about 70 epoch
训练词语: testtravel2
名称: strangeResult_v20_trainingData.zip
大小 (KB): 14491
类型: Training Data
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: testtravel2.pt
大小 (KB): 4
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
