Replace tokenizer.json with Xenova M2M-100 418M version (correct SP-aligned vocab IDs) for SMaLL-100

Fix tokenizer.json: add language tokens to added_tokens for alirezamsh/small100

Fix tokenizer.json for SMaLL-100 (correct BPE merge order)

Fix tokenizer.json merge order for SMaLL-100 INT8

Fix tokenizer.json merge order for SMaLL-100 INT8

Add tokenizer.json for SMaLL-100 INT8

Upload small100 ONNX INT8