Undesired output texts
avec ilaniobDetailsDocPorscheZGTUsBudAirIVEAirAIRddairperwarb11ixiabilityfertilmeni AcidBobFirstinariaÖhighbridgeiędzyAgmeniicaKa})+\X
In past, my prompt was giving proper human readable text but now above is what this model is outputing. I don't know what is wrong
Hi @Waisullah ,
Possible reasons for the issue:
Using inference settings like temperature=0, top_p=1.0, or repetition_penalty=1.0 might lead to flat, repetitive, or overly deterministic responses.
If the model weights or tokenizer files were only partially downloaded or got corrupted during caching, the model may start behaving unpredictably.
Running a quantized version of the model can sometimes lead to inaccurate or irrelevant outputs.
Also, could you please refer to the this link.
Thank you.