Create README.md

README.md

@@ -0,0 +1,47 @@
+---
+tags:
+- translation
+- torch==1.8.0
+---
+### marianmt-zh_cn-th
+* source languages: zh_cn
+* target languages: th
+* dataset: 
+* model: transformer-align
+* pre-processing: normalization + SentencePiece
+* test set translations: 
+* test set scores: 
+
+## Training
+
+```
+export WANDB_PROJECT=marianmt-zh_cn-th
+python train_model.py --input_fname ../data/v1/Train.csv \
+	--output_dir ../models/marianmt-zh_cn-th \
+	--source_lang zh --target_lang th \
+	--metric_tokenize th_syllable --fp16 --batch_size 64
+```
+
+## Usage
+
+```
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+ 
+tokenizer = AutoTokenizer.from_pretrained("cstorm125/marianmt-zh_cn-th")
+model = AutoModelForSeq2SeqLM.from_pretrained("cstorm125/marianmt-zh_cn-th").cpu()
+
+src_text = [
+    '我爱你',
+    '我想吃米饭',
+]
+translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
+print([tokenizer.decode(t, skip_special_tokens=True) for t in translated])
+
+> ['ผมรักคุณนะ', 'ฉันอยากกินข้าว']
+```
+
+## Requirements
+```
+transformers==4.6.0
+torch==1.8.0
+```