TheBloke commited on
Commit
f67f655
1 Parent(s): ddb010e

GPTQ model commit

Browse files
config.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/workspace/process/nondzu_mistral-7b-codealpaca-lora/source",
3
+ "architectures": [
4
+ "MistralForCausalLM"
5
+ ],
6
+ "bos_token_id": 1,
7
+ "eos_token_id": 2,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 4096,
10
+ "initializer_range": 0.02,
11
+ "intermediate_size": 14336,
12
+ "max_position_embeddings": 32768,
13
+ "model_type": "mistral",
14
+ "num_attention_heads": 32,
15
+ "num_hidden_layers": 32,
16
+ "num_key_value_heads": 8,
17
+ "pad_token_id": 0,
18
+ "pretraining_tp": 1,
19
+ "quantization_config": {
20
+ "batch_size": 1,
21
+ "bits": 4,
22
+ "block_name_to_quantize": "model.layers",
23
+ "damp_percent": 0.1,
24
+ "desc_act": true,
25
+ "disable_exllama": false,
26
+ "group_size": 128,
27
+ "max_input_length": null,
28
+ "model_seqlen": 4096,
29
+ "module_name_preceding_first_block": [
30
+ "model.embed_tokens"
31
+ ],
32
+ "pad_token_id": null,
33
+ "quant_method": "gptq",
34
+ "sym": true,
35
+ "tokenizer": null,
36
+ "true_sequential": true,
37
+ "use_cuda_fp16": true
38
+ },
39
+ "rms_norm_eps": 1e-05,
40
+ "rope_theta": 10000.0,
41
+ "sliding_window": 4096,
42
+ "tie_word_embeddings": false,
43
+ "torch_dtype": "float16",
44
+ "transformers_version": "4.34.1",
45
+ "use_cache": true,
46
+ "vocab_size": 32000
47
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.35.0.dev0"
6
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c2261bdbc1878c1ba9944ec43007a9d605f68dc22902b06e191994d962eea47
3
+ size 4158662096
quantize_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bits": 4,
3
+ "group_size": 128,
4
+ "damp_percent": 0.01,
5
+ "desc_act": true,
6
+ "sym": true,
7
+ "true_sequential": true
8
+ }
run-humaneval.py ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import torch
2
+ import torch.distributed as dist
3
+ import torch.multiprocessing as mp
4
+ from transformers import AutoTokenizer, LlamaForCausalLM
5
+ from torch.nn.parallel import DistributedDataParallel as DDP
6
+ from evalplus.data import get_human_eval_plus, write_jsonl
7
+ import os
8
+ from tqdm import tqdm # import tqdm
9
+
10
+ def setup(rank, world_size):
11
+ os.environ['MASTER_ADDR'] = 'localhost'
12
+ os.environ['MASTER_PORT'] = '12355'
13
+ dist.init_process_group("gloo", rank=rank, world_size=world_size)
14
+
15
+ def cleanup():
16
+ dist.destroy_process_group()
17
+
18
+ def generate_one_completion(ddp_model, tokenizer, prompt: str):
19
+ tokenizer.pad_token = tokenizer.eos_token
20
+ inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=4096)
21
+
22
+ # Generate
23
+ generate_ids = ddp_model.module.generate(inputs.input_ids.to("cuda"), max_new_tokens=384, do_sample=True, top_p=0.75, top_k=40, temperature=0.1, pad_token_id=tokenizer.eos_token_id)
24
+ completion = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
25
+ completion = completion.replace(prompt, "").split("\n\n\n")[0]
26
+
27
+ print("-------------------")
28
+ print(completion)
29
+ return completion
30
+
31
+ def run(rank, world_size):
32
+ setup(rank, world_size)
33
+
34
+ model_path = "Nondzu/Mistral-7B-codealpaca-lora"
35
+ model = LlamaForCausalLM.from_pretrained(model_path,load_in_8bit=True)
36
+ ddp_model = DDP(model, device_ids=[rank])
37
+ tokenizer = AutoTokenizer.from_pretrained(model_path)
38
+
39
+ problems = get_human_eval_plus()
40
+ num_samples_per_task = 1
41
+
42
+ samples = [
43
+ dict(task_id=task_id, completion=generate_one_completion(ddp_model, tokenizer, problems[task_id]["prompt"]))
44
+ for task_id in tqdm(problems) # add tqdm here
45
+ for _ in range(num_samples_per_task)
46
+ ]
47
+ write_jsonl(f"samples-Nondzu-Mistral-7B-codealpaca-lora-rank{rank}.jsonl", samples)
48
+
49
+ cleanup()
50
+
51
+ def main():
52
+ world_size = 1
53
+ mp.spawn(run, args=(world_size,), nprocs=world_size, join=True)
54
+
55
+ if __name__=="__main__":
56
+ main()
special_tokens_map.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<unk>",
4
+ "<s>",
5
+ "</s>"
6
+ ],
7
+ "bos_token": "<s>",
8
+ "eos_token": "</s>",
9
+ "pad_token": "</s>",
10
+ "unk_token": "<unk>"
11
+ }
tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055
3
+ size 493443
tokenizer_config.json ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
+ "added_tokens_decoder": {
5
+ "0": {
6
+ "content": "<unk>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "1": {
14
+ "content": "<s>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "2": {
22
+ "content": "</s>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ }
29
+ },
30
+ "additional_special_tokens": [
31
+ "<unk>",
32
+ "<s>",
33
+ "</s>"
34
+ ],
35
+ "bos_token": "<s>",
36
+ "chat_template": "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token + ' ' }}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
37
+ "clean_up_tokenization_spaces": false,
38
+ "eos_token": "</s>",
39
+ "legacy": true,
40
+ "model_max_length": 1000000000000000019884624838656,
41
+ "pad_token": "</s>",
42
+ "sp_model_kwargs": {},
43
+ "spaces_between_special_tokens": false,
44
+ "tokenizer_class": "LlamaTokenizer",
45
+ "tokenizer_file": "/home/kamil/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.1/snapshots/7ad5799710574ba1c1d953eba3077af582f3a773/tokenizer.json",
46
+ "trust_remote_code": false,
47
+ "unk_token": "<unk>",
48
+ "use_default_system_prompt": true,
49
+ "use_fast": true
50
+ }