microsoft
/

Phi-3.5-mini-instruct

Text Generation

text-generation-inference

Model card Files Files and versions

Resources

View closed (15)

Updated modelling.py - replaced deprecated get_max_length() with get_max_cache_shape() for transformers 4.49+

#41 opened 2 months ago by

Feedback on Hungarian Language Support

#40 opened 4 months ago by

megvadulthangya

fix bug 'DynamicCache' object has no attribute 'get_max_length' , updata modeling_phi3.py

#39 opened 5 months ago by

no need for custom_code anymore

#38 opened 5 months ago by

Update modeling_phi3.py for compatibility with transformers 4.49

#35 opened 6 months ago by

Fine tuning phi-3.5 or phi-2 ?

#33 opened 6 months ago by

Request: DOI

#32 opened 7 months ago by

I have 3 apples. Last week I ate 2 apples. How many apples do I have?

#31 opened 7 months ago by

Model not responding

#30 opened 7 months ago by

Phi-3.5-mini-instruct: CPU vs. GPU Mismatch Error in modeling_phi3.py During Generation

#29 opened 7 months ago by

Determinism Challenges with Microsoft Phi-3.5-mini-instruct Across Different GPU Architectures

#28 opened 8 months ago by

microsoft/Phi-3.5-mini-instruct not working with FA2 due to position_ids

#27 opened 8 months ago by

Looking forward to the phi 4 model release.

#26 opened 9 months ago by

Hallucinated reply if you prompt same prompt several times. Ollama and phi3.5

#25 opened 9 months ago by

curl run on colab

#23 opened 11 months ago by

KeyError: 'phi3' when attempting to train in Oogabooga

#21 opened 11 months ago by

Please support phi-3.5-mini-instruct in llama.cpp

#20 opened 12 months ago by

Can't install flash_attn package on WSL2

#19 opened 12 months ago by

How to add new token or special token?

#16 opened 12 months ago by

Evaluation of Phi-3.5 on long-context BABILong bench

#12 opened about 1 year ago by

Phi3ForCausalLM.forward() got an unexpected keyword argument 'decoder_input_ids'

#11 opened about 1 year ago by

tokenizer.model_max_length=2048 in sample_finetune.py

#10 opened about 1 year ago by

Very safe model!

#6 opened about 1 year ago by

SicariusSicariiStuff

Will there be int 4 ONNX DirectML versions in the future?

#5 opened about 1 year ago by

My quants and my "silly" version.

#4 opened about 1 year ago by

Why not include MedQA in your benchmarks?

#1 opened about 1 year ago by