metadata

language:
  - en
library_name: transformers
pipeline_tag: text-generation
datasets:
  - jondurbin/airoboros-2.2
  - Open-Orca/OpenOrca
  - garage-bAInd/Open-Platypus
  - WizardLM/WizardLM_evol_instruct_V2_196k
tags:
  - llama-2
  - code
license: llama2
model-index:
  - name: SpeechlessCoder
    results:
      - task:
          type: text-generation
        dataset:
          type: openai_humaneval
          name: HumanEval
        metrics:
          - name: pass@1
            type: pass@1
            value: 75.61
            verified: false

speechless-codellama-34b-v2.0

Code: https://github.com/uukuguy/speechless

Use the following datasets to fine-tune codellama/CodeLlama-34B in order to improve the model's inference and planning capabilities.

Total 153,013 samples.

jondurbin/airoboros-2.2: Filter categories related to coding, reasoning and planning. 23,462 samples.
Open-Orca/OpenOrca: Filter the 'cot' category in 1M GPT4 dataset. 74,440 samples.
garage-bAInd/Open-Platypus: 100%, 24,926 samples.
WizardLM/WizardLM_evol_instruct_V2_196k: Coding coversation part. 30,185 samples

How to Prompt the Model

This model accepts the Alpaca instruction format.

For example:

You are an intelligent programming assistant.

### Instruction:
Implement a linked list in C++

### Response:

HumanEval

human-eval	pass@1
humaneval-python	75.61

Big Code Models Leaderboard

Models	pass@1
Phind-CodeLlama-34B-v2	71.95
WizardCoder-Python-34B-V1.0	70.73
Phind-CodeLlama-34B-Python-v1	70.22
Phind-CodeLlama-34B-v1	65.85
WizardCoder-Python-13B-V1.0	62.19
WizardCoder-15B-V1.0	58.12
CodeLlama-34B-Python	53.29
CodeLlama-34B-Instruct	50.79
CodeLlama-13B-Instruct	50.6
CodeLlama-34B	45.11
CodeLlama-13B-Python	42.89
CodeLlama-13B	35.07

NL2SQL

SQL-EVAL: 125/175 (71.43%)

Average rate of exact match: 67.43%

Average correct rate: 71.43%

GPT4: 130/175 (74.29%)
GPT3-Turbo-0613: 105/174 (60.00%)

lm-evaluation-harness

Open LLM Leaderboard

Metric	Value
ARC	54.35
HellaSwag	75.65
MMLU	54.67
TruthfulQA	45.21
Average	57.47

H800-80G x 2

transformers=4.33.0

flash-attn=2.1.0

bitsandbytes=0.41.1

peft=0.5.0

Training Arguments


lr	2e-4
lr_scheduler_type	cosine
weight_decay	0.0
optim	paged_adamw_8bit
flash_attention	True
rerope	False
max_new_tokens	8192
num_train_epochs	3
bits	4
lora_r	64
lora_alpha	16
lora_dropout	0.05
double_quant	True
quant_type	nf4
dataset_format	airoboros
mini_batch_size	4
grandient_accumulation_steps	16
bf16	True


epoch	3.0
etrain_loss	0.4261
etrain_runtime	1 day, 14:42:57.87
etrain_samples_per_second	3.227
etrain_steps_per_second	0.025
eeval_loss	0.4537
eeval_runtime	0:00:36.19
eeval_samples_per_second	5.525
eeval_steps_per_second	2.763

Code Llama

Code Llama is a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 34 billion parameters. This is the repository for the base 13B version in the Hugging Face Transformers format. This model is designed for general code synthesis and understanding. Links to other models can be found in the index at the bottom.

	Base Model	Python	Instruct
7B	codellama/CodeLlama-7b-hf	codellama/CodeLlama-7b-Python-hf	codellama/CodeLlama-7b-Instruct-hf
13B	codellama/CodeLlama-13b-hf	codellama/CodeLlama-13b-Python-hf	codellama/CodeLlama-13b-Instruct-hf
34B	codellama/CodeLlama-34b-hf	codellama/CodeLlama-34b-Python-hf	codellama/CodeLlama-34b-Instruct-hf

Model Use

To use this model, please make sure to install transformers from main until the next version is released:

pip install git+https://github.com/huggingface/transformers.git@main accelerate

Model capabilities:

Code completion.
Infilling.
Instructions / chat.
Python specialist.

from transformers import AutoTokenizer
import transformers
import torch

model = "codellama/CodeLlama-13b-hf"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

sequences = pipeline(
    'import socket\n\ndef ping_exponential_backoff(host: str):',
    do_sample=True,
    top_k=10,
    temperature=0.1,
    top_p=0.95,
    num_return_sequences=1,
    eos_token_id=tokenizer.eos_token_id,
    max_length=200,
)
for seq in sequences:
    print(f"Result: {seq['generated_text']}")

Model Details

*Note: Use of this model is governed by the Meta license. Meta developed and publicly released the Code Llama family of large language models (LLMs).

Model Developers Meta

Variations Code Llama comes in three model sizes, and three variants:

Code Llama: base models designed for general code synthesis and understanding
Code Llama - Python: designed specifically for Python
Code Llama - Instruct: for instruction following and safer deployment

All variants are available in sizes of 7B, 13B and 34B parameters.

This repository contains the base version of the 13B parameters model.

Input Models input text only.

Output Models generate text only.

Model Architecture Code Llama is an auto-regressive language model that uses an optimized transformer architecture.

Model Dates Code Llama and its variants have been trained between January 2023 and July 2023.

Status This is a static model trained on an offline dataset. Future versions of Code Llama - Instruct will be released as we improve model safety with community feedback.

License A custom commercial license is available at: https://ai.meta.com/resources/models-and-libraries/llama-downloads/

Research Paper More information can be found in the paper "Code Llama: Open Foundation Models for Code" or its arXiv page.

Intended Use

Intended Use Cases Code Llama and its variants is intended for commercial and research use in English and relevant programming languages. The base model Code Llama can be adapted for a variety of code synthesis and understanding tasks, Code Llama - Python is designed specifically to handle the Python programming language, and Code Llama - Instruct is intended to be safer to use for code assistant and generation applications.

Out-of-Scope Uses Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in languages other than English. Use in any other way that is prohibited by the Acceptable Use Policy and Licensing Agreement for Code Llama and its variants.

Hardware and Software

Training Factors We used custom training libraries. The training and fine-tuning of the released models have been performed Meta’s Research Super Cluster.

Carbon Footprint In aggregate, training all 9 Code Llama models required 400K GPU hours of computation on hardware of type A100-80GB (TDP of 350-400W). Estimated total emissions were 65.3 tCO2eq, 100% of which were offset by Meta’s sustainability program.

Training Data

All experiments reported here and the released models have been trained and fine-tuned using the same data as Llama 2 with different weights (see Section 2 and Table 1 in the research paper for details).

Evaluation Results

See evaluations for the main models and detailed ablations in Section 3 and safety evaluations in Section 4 of the research paper.

Ethical Considerations and Limitations

Code Llama and its variants are a new technology that carries risks with use. Testing conducted to date has been in English, and has not covered, nor could it cover all scenarios. For these reasons, as with all LLMs, Code Llama’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate or objectionable responses to user prompts. Therefore, before deploying any applications of Code Llama, developers should perform safety testing and tuning tailored to their specific applications of the model.

Please see the Responsible Use Guide available available at https://ai.meta.com/llama/responsible-user-guide.

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	50.96
ARC (25-shot)	54.35
HellaSwag (10-shot)	75.65
MMLU (5-shot)	54.67
TruthfulQA (0-shot)	45.21
Winogrande (5-shot)	73.56
GSM8K (5-shot)	11.6
DROP (3-shot)	41.71