What are the hardware requirements for CodeLlama 34B Python Fp16?

CodeLlama 34B Python Fp16 requires approximately 67.5 GB of VRAM and supports a context window of 16K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed CodeLlama 34B Python Fp16 and how large is it?

CodeLlama 34B Python Fp16 is developed by TheBloke, a model with 34b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does CodeLlama 34B Python Fp16 perform on standard benchmarks?

CodeLlama 34B Python Fp16 has the following published scores: MMLU 32.95, HumanEval (Python) 53.29. Compare against reference models on this page or on the LLM Explorer leaderboards.

CodeLlama 34B Python Fp16 by TheBloke — VRAM 67.5GB, 16K context

Name: CodeLlama 34B Python Fp16
Author: TheBloke

CodeLlama 34B Python Fp16 is an open-source language model by TheBloke. Features: 34b LLM, VRAM: 67.5GB, Context: 16K, License: llama2, Quantized, Code Generating, LLM Explorer Score: 0.12, Arc: 38.1, HellaSwag: 34.8, MMLU: 33, GSM8K: 20, HumanEval: 53.3.

Codegen Codellama Custom code Endpoints compatible Fp16 Llama Llama2 Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/TheBloke/CodeLlama-34B-Python-fp16

CodeLlama 34B Python Fp16 Parameters and Internals

Model Type

text generation, code synthesis

Use Cases

Areas:

Commercial, Research

Applications:

Code synthesis, Code understanding, Python specific tasks

Primary Use Cases:

Commercial code generation, Research projects in programming and AI

Limitations:

Not for non-English languages, Should not violate laws or regulations

Considerations:

Adhere to legal compliances.

Additional Notes

Intended for Python code tasks; requires specific handling for different operations.

Supported Languages

English (proficient), Python (proficient)

Training Details

Data Sources:

Same data as Llama 2 with different weights

Methodology:

Fine-tuned with additional instruction data for CodeLlama - Instruct

Context Length:

100000

Training Time:

January 2023 - July 2023

Hardware Used:

Meta’s Research Super Cluster, A100-80GB GPUs

Model Architecture:

Autoregressive language model using optimized transformer architectures

Responsible Ai Considerations

Fairness:

Tested primarily in English.

Transparency:

Model's potential outputs cannot be predicted in advance.

Accountability:

Developers should perform safety testing and tuning tailored to their specific applications before deployment.

Mitigation Strategies:

Please see the Responsible Use Guide for more details.

Input Output

Input Format:

Text only

Accepted Modalities:

text

Output Format:

Text only

Performance Tips:

Use 'trust_remote_code=True' for optimal operation due to a change in the RoPE Theta value

LLM Name	CodeLlama 34B Python Fp16
Repository 🤗	https://huggingface.co/TheBloke/CodeLlama-34B-Python-fp16
Base Model(s)	CodeLlama 34B Python Hf codellama/CodeLlama-34b-Python-hf
Model Size	34b
Required VRAM	67.5 GB
Updated	2026-07-26
Maintainer	TheBloke
Model Type	llama
Model Files	9.8 GB: 1-of-7 9.7 GB: 2-of-7 9.7 GB: 3-of-7 9.7 GB: 4-of-7 9.7 GB: 5-of-7 9.7 GB: 6-of-7 9.2 GB: 7-of-7
Quantization Type	fp16
Generates Code	Yes
Model Architecture	LlamaForCausalLM
License	llama2
Context Length	16384
Model Max Length	16384
Transformers Version	4.32.0
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	float16

Rank the CodeLlama 34B Python Fp16 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to CodeLlama 34B Python Fp16

Best Alternatives	Context / RAM	Downloads	Likes
CodeLlama 34B Instruct Fp16	16K / 67.5 GB	131	7
CodeLlama 34B Instruct Hf 4bit	16K / 19.4 GB	63	2
Phind Codellama 34B V2 EXL2	16K / GB	8	16
Codellama Extraction	16K / 67.6 GB	6	0
CodeLlama 34B Fp16	16K / 67.5 GB	25	4
Codellama 34B Bnb 4bit	16K / 18.2 GB	38	4
XwinCoder 34B 4.0bpw H6 EXL2	16K / 17.4 GB	3	1
...Codellama 34B V2 Megacode EXL2	16K / GB	3	10
...gpt 32K Codellama 34B Instruct	32K / 67.5 GB	56	2
CodeLlama 34B Instruct Hf	16K / 67.5 GB	22872	304

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/CodeLlama-34B-Python-fp16.