Ceyhun Derinbogaz commited on
Commit
9daee60
1 Parent(s): 12b23b9

initial commit

Browse files
README.md CHANGED
@@ -1,3 +1,61 @@
1
  ---
2
  license: bsd-3-clause
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: bsd-3-clause
3
  ---
4
+ # CodeGen (CodeGen-Mono 350M)
5
+ This is a clone of CodeGen project which is optimized to run on CPU by using the ONNX optimisations.
6
+ The reason we created ONNX version of the original version is, we wanted to make it possible for ICortex kernel users
7
+ to easily generate code for their use case without using a GPU.
8
+ Original model can be found [here](https://huggingface.co/Salesforce/codegen-350M-mono).
9
+
10
+ ## Model description
11
+
12
+ CodeGen is a family of autoregressive language models for **program synthesis** from the paper: [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong. The models are originally released in [this repository](https://github.com/salesforce/CodeGen), under 3 pre-training data variants (`NL`, `Multi`, `Mono`) and 4 model size variants (`350M`, `2B`, `6B`, `16B`).
13
+
14
+ The checkpoint included in this repository is denoted as **CodeGen-Mono 350M** in the paper, where "Mono" means the model is initialized with *CodeGen-Multi 350M* and further pre-trained on a Python programming language dataset, and "350M" refers to the number of trainable parameters.
15
+
16
+ ## Training data
17
+
18
+ This checkpoint (CodeGen-Mono 350M) was firstly initialized with *CodeGen-Multi 350M*, and then pre-trained on BigPython dataset. The data consists of 71.7B tokens of Python programming language. See Section 2.1 of the [paper](https://arxiv.org/abs/2203.13474) for more details.
19
+
20
+ ## Training procedure
21
+
22
+ CodeGen was trained using cross-entropy loss to maximize the likelihood of sequential inputs.
23
+ The family of models are trained using multiple TPU-v4-512 by Google, leveraging data and model parallelism.
24
+ See Section 2.3 of the [paper](https://arxiv.org/abs/2203.13474) for more details.
25
+
26
+ ## Evaluation results
27
+
28
+ We evaluate our models on two code generation benchmark: HumanEval and MTPB. Please refer to the [paper](https://arxiv.org/abs/2203.13474) for more details.
29
+
30
+
31
+ ## Intended Use and Limitations
32
+
33
+ As an autoregressive language model, CodeGen is capable of extracting features from given natural language and programming language texts, and calculating the likelihood of them.
34
+ However, the model is intended for and best at **program synthesis**, that is, generating executable code given English prompts, where the prompts should be in the form of a comment string. The model can complete partially-generated code as well.
35
+
36
+ ## How to use
37
+
38
+ This model can be easily loaded using the `AutoModelForCausalLM` functionality:
39
+
40
+ ```python
41
+ from transformers import AutoTokenizer, AutoModelForCausalLM
42
+ tokenizer = AutoTokenizer.from_pretrained("Salesforce/codegen-350M-mono")
43
+ model = AutoModelForCausalLM.from_pretrained("Salesforce/codegen-350M-mono")
44
+
45
+ text = "def hello_world():"
46
+ input_ids = tokenizer(text, return_tensors="pt").input_ids
47
+
48
+ generated_ids = model.generate(input_ids, max_length=128)
49
+ print(tokenizer.decode(generated_ids[0], skip_special_tokens=True))
50
+ ```
51
+
52
+ ## BibTeX entry and citation info
53
+
54
+ ```bibtex
55
+ @article{Nijkamp2022ACP,
56
+ title={A Conversational Paradigm for Program Synthesis},
57
+ author={Nijkamp, Erik and Pang, Bo and Hayashi, Hiroaki and Tu, Lifu and Wang, Huan and Zhou, Yingbo and Savarese, Silvio and Xiong, Caiming},
58
+ journal={arXiv preprint},
59
+ year={2022}
60
+ }
61
+ ```
added_tokens.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "\t\t": 50294,
3
+ "\t\t\t": 50293,
4
+ "\t\t\t\t": 50292,
5
+ "\t\t\t\t\t": 50291,
6
+ "\t\t\t\t\t\t": 50290,
7
+ "\t\t\t\t\t\t\t": 50289,
8
+ "\t\t\t\t\t\t\t\t": 50288,
9
+ "\t\t\t\t\t\t\t\t\t": 50287,
10
+ " ": 50286,
11
+ " ": 50285,
12
+ " ": 50284,
13
+ " ": 50283,
14
+ " ": 50282,
15
+ " ": 50281,
16
+ " ": 50280,
17
+ " ": 50279,
18
+ " ": 50278,
19
+ " ": 50277,
20
+ " ": 50276,
21
+ " ": 50275,
22
+ " ": 50274,
23
+ " ": 50273,
24
+ " ": 50272,
25
+ " ": 50271,
26
+ " ": 50270,
27
+ " ": 50269,
28
+ " ": 50268,
29
+ " ": 50267,
30
+ " ": 50266,
31
+ " ": 50265,
32
+ " ": 50264,
33
+ " ": 50263,
34
+ " ": 50262,
35
+ " ": 50261,
36
+ " ": 50260,
37
+ " ": 50259,
38
+ " ": 50258,
39
+ " ": 50257
40
+ }
config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Salesforce/codegen-350M-mono",
3
+ "activation_function": "gelu_new",
4
+ "architectures": [
5
+ "CodeGenForCausalLM"
6
+ ],
7
+ "attn_pdrop": 0.0,
8
+ "bos_token_id": 1,
9
+ "embd_pdrop": 0.0,
10
+ "eos_token_id": 2,
11
+ "gradient_checkpointing": false,
12
+ "initializer_range": 0.02,
13
+ "layer_norm_epsilon": 1e-05,
14
+ "n_ctx": 2048,
15
+ "n_embd": 1024,
16
+ "n_head": 16,
17
+ "n_inner": null,
18
+ "n_layer": 20,
19
+ "n_positions": 2048,
20
+ "pad_token_id": 0,
21
+ "resid_pdrop": 0.0,
22
+ "rotary_dim": 32,
23
+ "scale_attn_weights": true,
24
+ "summary_activation": null,
25
+ "summary_first_dropout": 0.1,
26
+ "summary_proj_to_labels": true,
27
+ "summary_type": "cls_index",
28
+ "summary_use_proj": true,
29
+ "task_specific_params": {
30
+ "text-generation": {
31
+ "do_sample": true,
32
+ "max_length": 50,
33
+ "temperature": 1.0
34
+ }
35
+ },
36
+ "tie_word_embeddings": false,
37
+ "tokenizer_class": "GPT2Tokenizer",
38
+ "torch_dtype": "float16",
39
+ "transformers_version": "4.21.3",
40
+ "use_cache": false,
41
+ "vocab_size": 51200
42
+ }
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee98aa9903134b586986103601652e734249e9f5fb996d48dc73a3d4689f29ad
3
+ size 1431735160
special_tokens_map.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<|endoftext|>",
3
+ "eos_token": "<|endoftext|>",
4
+ "unk_token": "<|endoftext|>"
5
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "bos_token": "<|endoftext|>",
4
+ "eos_token": "<|endoftext|>",
5
+ "model_max_length": 1024,
6
+ "name_or_path": "Salesforce/codegen-350M-mono",
7
+ "special_tokens_map_file": null,
8
+ "tokenizer_class": "CodeGenTokenizer",
9
+ "unk_token": "<|endoftext|>"
10
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff