Declare base_model_relation: the weights are converted, not fine-tuned

Add upstream attribution notice

Add Apache-2.0 license text

Add config.json

Add five-task Jev comparison reports

Model card: guardrails 0.808

benchmark-coreml: post tokenizer fix

Model card: e8 buckets

Add e8 + compression + ANE profile reports

Add laya_multilingual_e8_L1024_options32 (int8 embedding table)

Add laya_multilingual_e8_L512_options32 (int8 embedding table)

Add laya_multilingual_e8_L256_options32 (int8 embedding table)

Add laya_multilingual_e8_L128_options32 (int8 embedding table)

Model card: FluidUse

Model card: benchmark table

Add accuracy benchmark reports

Model card: L1024

Upload reports/verification-multilingual-L1024.json with huggingface_hub

Add L1024 bucket

Add verification reports

Add model card

Add tokenizer.json

Add laya_multilingual_fp16_L512_options32

Add laya_multilingual_fp16_L256_options32

Add laya_multilingual_fp16_L128_options32

initial commit