Trimmed google/mt5-base for Thai (32768 tokens)
Browse files- README.md +24 -0
- config.json +28 -0
- generation_config.json +7 -0
- model.safetensors +3 -0
- special_tokens_map.json +6 -0
- spiece.model +3 -0
- tokenizer_config.json +6 -0
README.md
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language: tha
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
tags: [trimmed, mt5, seq2seq]
|
| 5 |
+
base_model: google/mt5-base
|
| 6 |
+
datasets:
|
| 7 |
+
- Lumberjackk/fineweb-2-trimming
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# mt5-base-tha-32768
|
| 11 |
+
|
| 12 |
+
Version de [google/mt5-base](https://huggingface.co/google/mt5-base) avec vocabulaire réduit pour **Thai**.
|
| 13 |
+
|
| 14 |
+
| | Original | Trimmed |
|
| 15 |
+
|---|---|---|
|
| 16 |
+
| Vocabulaire | 250,100 | 32,768 |
|
| 17 |
+
| Paramètres | 582,401,280 | 248,560,896 |
|
| 18 |
+
|
| 19 |
+
## Usage
|
| 20 |
+
```python
|
| 21 |
+
from transformers import T5Tokenizer, AutoModelForSeq2SeqLM
|
| 22 |
+
tokenizer = T5Tokenizer.from_pretrained("lbourdois/mt5-base-tha-32768")
|
| 23 |
+
model = AutoModelForSeq2SeqLM.from_pretrained("lbourdois/mt5-base-tha-32768")
|
| 24 |
+
```
|
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/home/patrick/hugging_face/t5/mt5-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"MT5ForConditionalGeneration"
|
| 5 |
+
],
|
| 6 |
+
"d_ff": 2048,
|
| 7 |
+
"d_kv": 64,
|
| 8 |
+
"d_model": 768,
|
| 9 |
+
"decoder_start_token_id": 0,
|
| 10 |
+
"dropout_rate": 0.1,
|
| 11 |
+
"eos_token_id": 1,
|
| 12 |
+
"feed_forward_proj": "gated-gelu",
|
| 13 |
+
"initializer_factor": 1.0,
|
| 14 |
+
"is_encoder_decoder": true,
|
| 15 |
+
"layer_norm_epsilon": 1e-06,
|
| 16 |
+
"model_type": "mt5",
|
| 17 |
+
"num_decoder_layers": 12,
|
| 18 |
+
"num_heads": 12,
|
| 19 |
+
"num_layers": 12,
|
| 20 |
+
"output_past": true,
|
| 21 |
+
"pad_token_id": 0,
|
| 22 |
+
"relative_attention_num_buckets": 32,
|
| 23 |
+
"tie_word_embeddings": false,
|
| 24 |
+
"tokenizer_class": "T5Tokenizer",
|
| 25 |
+
"transformers_version": "4.10.0.dev0",
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 32768
|
| 28 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "5.3.0.dev0"
|
| 7 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a5b4f535f5d8d7e4afc5ea6e549f4bc169f01110bceaf74b0c8cfdb7799fb705
|
| 3 |
+
size 994277224
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eos_token": "</s>",
|
| 3 |
+
"unk_token": "<unk>",
|
| 4 |
+
"pad_token": "<pad>",
|
| 5 |
+
"additional_special_tokens": []
|
| 6 |
+
}
|
spiece.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af7429a49840b763f47c0d582103943285e2ca582cba1c1fd2db0a5ef0e6dda7
|
| 3 |
+
size 763620
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"eos_token": "</s>",
|
| 3 |
+
"unk_token": "<unk>",
|
| 4 |
+
"pad_token": "<pad>",
|
| 5 |
+
"extra_ids": 0
|
| 6 |
+
}
|