Locutusque
commited on
Commit
•
af9da65
1
Parent(s):
0a74920
Upload 11 files
Browse files- adapter_config.json +20 -0
- added_tokens.json +5 -0
- config.json +39 -0
- generation_config.json +6 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- run_model.py +38 -0
- special_tokens_map.json +11 -0
- tokenizer.json +0 -0
- tokenizer_config.json +9 -0
- vocab.json +0 -0
adapter_config.json
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"base_model_name_or_path": "EleutherAI/gpt-j-6b",
|
3 |
+
"bias": "none",
|
4 |
+
"fan_in_fan_out": false,
|
5 |
+
"inference_mode": true,
|
6 |
+
"init_lora_weights": true,
|
7 |
+
"layers_pattern": null,
|
8 |
+
"layers_to_transform": null,
|
9 |
+
"lora_alpha": 32,
|
10 |
+
"lora_dropout": 0.1,
|
11 |
+
"modules_to_save": null,
|
12 |
+
"peft_type": "LORA",
|
13 |
+
"r": 16,
|
14 |
+
"revision": null,
|
15 |
+
"target_modules": [
|
16 |
+
"q_proj",
|
17 |
+
"v_proj"
|
18 |
+
],
|
19 |
+
"task_type": "CAUSAL_LM"
|
20 |
+
}
|
added_tokens.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"<|ASSISTANT|>": 50257,
|
3 |
+
"<|SYSTEM|>": 50259,
|
4 |
+
"<|USER|>": 50258
|
5 |
+
}
|
config.json
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "gpt2-large",
|
3 |
+
"activation_function": "gelu_new",
|
4 |
+
"architectures": [
|
5 |
+
"GPT2LMHeadModel"
|
6 |
+
],
|
7 |
+
"attn_pdrop": 0.1,
|
8 |
+
"bos_token_id": 50256,
|
9 |
+
"embd_pdrop": 0.1,
|
10 |
+
"eos_token_id": 50256,
|
11 |
+
"initializer_range": 0.02,
|
12 |
+
"layer_norm_epsilon": 1e-05,
|
13 |
+
"model_type": "gpt2",
|
14 |
+
"n_ctx": 1024,
|
15 |
+
"n_embd": 1280,
|
16 |
+
"n_head": 20,
|
17 |
+
"n_inner": null,
|
18 |
+
"n_layer": 36,
|
19 |
+
"n_positions": 1024,
|
20 |
+
"reorder_and_upcast_attn": false,
|
21 |
+
"resid_pdrop": 0.1,
|
22 |
+
"scale_attn_by_inverse_layer_idx": false,
|
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 |
+
}
|
34 |
+
},
|
35 |
+
"torch_dtype": "bfloat16",
|
36 |
+
"transformers_version": "4.30.1",
|
37 |
+
"use_cache": true,
|
38 |
+
"vocab_size": 50260
|
39 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 50256,
|
4 |
+
"eos_token_id": 50256,
|
5 |
+
"transformers_version": "4.30.1"
|
6 |
+
}
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bc74cffab094a7fb2e0ae37c1dcedc4a522700cf0909ee9e05d4b284f3437c0e
|
3 |
+
size 1548210205
|
run_model.py
ADDED
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
from transformers import GPT2Tokenizer, AutoModelForCausalLM
|
3 |
+
start_token = "<|ASSISTANT|>"
|
4 |
+
end_token = "<|"
|
5 |
+
tokenizer = GPT2Tokenizer.from_pretrained('gpt2-large')
|
6 |
+
model = AutoModelForCausalLM.from_pretrained('gpt2-large', torch_dtype=torch.bfloat16)
|
7 |
+
tokenizer.pad_token = "[PAD]"
|
8 |
+
tokenizer.eos_token = "<|endoftext|>"
|
9 |
+
tokenizer.add_special_tokens({"additional_special_tokens": ["<|ASSISTANT|>", "<|USER|>", "<|SYSTEM|>"]})
|
10 |
+
model.resize_token_embeddings(len(tokenizer))
|
11 |
+
model.load_state_dict(torch.load("/media/locutusque/T7/Projects/results/pytorch_model.bin"))
|
12 |
+
model.cuda()
|
13 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
14 |
+
def generate_text(model, tokenizer, prompt, max_length=1024):
|
15 |
+
prompt = f'<|USER|> {prompt} <|ASSISTANT|> '
|
16 |
+
input_ids = tokenizer.encode(prompt, add_special_tokens=True, return_tensors="pt").to(device)
|
17 |
+
attention_mask = torch.ones_like(input_ids).to(device)
|
18 |
+
output = model.generate(input_ids,
|
19 |
+
max_length=max_length,
|
20 |
+
do_sample=True,
|
21 |
+
top_k=0,
|
22 |
+
top_p=0.1,
|
23 |
+
temperature=0.75,
|
24 |
+
repetition_penalty=1.176,
|
25 |
+
pad_token_id=tokenizer.pad_token_id,
|
26 |
+
eos_token_id=tokenizer.eos_token_id,
|
27 |
+
attention_mask=attention_mask)
|
28 |
+
output_ids = tokenizer.decode(output[0], skip_special_tokens=False)
|
29 |
+
return output_ids
|
30 |
+
# Loop to interact with the model
|
31 |
+
while True:
|
32 |
+
prompt = input("Enter a prompt (or 'q' to quit): ")
|
33 |
+
if prompt == "q":
|
34 |
+
break
|
35 |
+
output_text = generate_text(model, tokenizer, prompt)
|
36 |
+
text_between_tokens = output_text[output_text.find(start_token) + len(start_token):]
|
37 |
+
out = text_between_tokens[:text_between_tokens.find(end_token)]
|
38 |
+
print(out)
|
special_tokens_map.json
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": [
|
3 |
+
"<|ASSISTANT|>",
|
4 |
+
"<|USER|>",
|
5 |
+
"<|SYSTEM|>"
|
6 |
+
],
|
7 |
+
"bos_token": "<|endoftext|>",
|
8 |
+
"eos_token": "<|endoftext|>",
|
9 |
+
"pad_token": "[PAD]",
|
10 |
+
"unk_token": "<|endoftext|>"
|
11 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"bos_token": "<|endoftext|>",
|
4 |
+
"clean_up_tokenization_spaces": true,
|
5 |
+
"eos_token": "<|endoftext|>",
|
6 |
+
"model_max_length": 1024,
|
7 |
+
"tokenizer_class": "GPT2Tokenizer",
|
8 |
+
"unk_token": "<|endoftext|>"
|
9 |
+
}
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|