keshan commited on
Commit
073ed17
1 Parent(s): 51d9287

initial preps with config and tokenizers

Browse files
.gitattributes CHANGED
@@ -14,3 +14,4 @@
14
  *.pb filter=lfs diff=lfs merge=lfs -text
15
  *.pt filter=lfs diff=lfs merge=lfs -text
16
  *.pth filter=lfs diff=lfs merge=lfs -text
 
 
14
  *.pb filter=lfs diff=lfs merge=lfs -text
15
  *.pt filter=lfs diff=lfs merge=lfs -text
16
  *.pth filter=lfs diff=lfs merge=lfs -text
17
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
config.py ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ from transformers import GPT2Config
2
+
3
+ model_dir = "sinhala-gpt2"
4
+
5
+ config = GPT2Config.from_pretrained("gpt2", resid_pdrop=0.0, embd_pdrop=0.0, attn_pdrop=0.0)
6
+ config.save_pretrained(model_dir)
7
+
run_clm_flax.py ADDED
@@ -0,0 +1 @@
 
 
1
+ /home/Keshan/transformers/examples/flax/language-modeling/run_clm_flax.py
si_tokenizer.py ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from datasets import load_dataset
2
+ from tokenizers import trainers, Tokenizer, normalizers, ByteLevelBPETokenizer
3
+
4
+ model_dir = "./sinhala-gpt2" # ${MODEL_DIR}
5
+
6
+ # load dataset
7
+ dataset = load_dataset("mc4", "si", split="train")
8
+
9
+ # Instantiate tokenizer
10
+ tokenizer = ByteLevelBPETokenizer()
11
+
12
+ def batch_iterator(batch_size=1000):
13
+ for i in range(0, len(dataset), batch_size):
14
+ yield dataset[i: i + batch_size]["text"]
15
+
16
+ # Customized training
17
+ tokenizer.train_from_iterator(batch_iterator(), vocab_size=50265, min_frequency=2, special_tokens=[
18
+ "<s>",
19
+ "<pad>",
20
+ "</s>",
21
+ "<unk>",
22
+ "<mask>",
23
+ ])
24
+
25
+ # Save files to disk
26
+ tokenizer.save(f"{model_dir}/tokenizer.json")
sinhala-gpt2/config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_function": "gelu_new",
3
+ "architectures": [
4
+ "GPT2LMHeadModel"
5
+ ],
6
+ "attn_pdrop": 0.0,
7
+ "bos_token_id": 50256,
8
+ "embd_pdrop": 0.0,
9
+ "eos_token_id": 50256,
10
+ "gradient_checkpointing": false,
11
+ "initializer_range": 0.02,
12
+ "layer_norm_epsilon": 1e-05,
13
+ "model_type": "gpt2",
14
+ "n_ctx": 1024,
15
+ "n_embd": 768,
16
+ "n_head": 12,
17
+ "n_inner": null,
18
+ "n_layer": 12,
19
+ "n_positions": 1024,
20
+ "resid_pdrop": 0.0,
21
+ "scale_attn_weights": true,
22
+ "summary_activation": null,
23
+ "summary_first_dropout": 0.1,
24
+ "summary_proj_to_labels": true,
25
+ "summary_type": "cls_index",
26
+ "summary_use_proj": true,
27
+ "task_specific_params": {
28
+ "text-generation": {
29
+ "do_sample": true,
30
+ "max_length": 50
31
+ }
32
+ },
33
+ "transformers_version": "4.9.0.dev0",
34
+ "use_cache": true,
35
+ "vocab_size": 50257
36
+ }
sinhala-gpt2/events.out.tfevents.1625744111.t1v-n-0b4afe6c-w-0.150137.3.v2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:91ed273fdad9323ae19e20ef9e232b1dfc8bf623ed07b83e375ccab2950b2c20
3
+ size 10560081
sinhala-gpt2/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff