NER, long audio, multilingual

[hududed]

@huseinzol05 was browsing huggingface for off-the-shelf models for NER and ASR for malay and southeast asian languages (e.g. english, indonesian, thai, singlish), found bert-bahasa-cased but that's no longer available, which led me here.

This work looks very promising thank you for making this available! I was wondering if you know of any models I can immediately use for NER (single or multilingual)? Or a translation model to english and then NER?
For ASR, I also have long audio >30 mins - which model could you recommend with higher accuracy?