LL

Llama 2 70b Hf

LlamaGGUF by mradermacherUpdated 2024-09-10
Download on Hugging Face
Parameters69B
Context4K
Layers80
Architecturellama
Smallest quant14.5 GB
Quick answer

Llama 2 70b Hf needs about 44.6 GB of memory at Q4_K_M (a 41.4 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 20 quants, from 14.5 GB (IQ1_S) to 48.8 GB (Q5_K_M).

Llama 2 70B HF converted to GGUF format

Llama 2 70B HF is a 69‑billion parameter autoregressive language model. It is converted to GGUF for use with llama.cpp. The model supports 4096 token context and is available in multiple quantizations.

Good at

  • Large‑scale text generation
  • High‑capacity language modeling
  • Long‑context handling

Limits

  • No vision support
  • License requires acceptance
  • Large VRAM needed for full precision

Best for

  • Research on large language models
  • Fine‑tuning for specialized tasks
  • Benchmarking inference performance

Who it’s for. Researchers and developers who need a 70B parameter model for text generation and want to run it locally with llama.cpp. The model is suitable for those with sufficient GPU memory and a need for high‑capacity language understanding.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S14.5 GB 16.4 GB
Too big Get file →
IQ1_M15.9 GB 17.8 GB
Too big Get file →
IQ2_XXS18.3 GB 20.1 GB
Too big Get file →
IQ2_XS20.3 GB 22.2 GB
Too big Get file →
IQ2_S21.4 GB 23.2 GB
Too big Get file →
IQ2_M23.2 GB 25.1 GB
Too big Get file →
Q2_K25.5 GB 27.3 GB
Too big Get file →
IQ3_XXS26.6 GB 28.4 GB
Too big Get file →
IQ3_XS28.3 GB 30.2 GB
Too big Get file →
IQ3_S29.9 GB 31.8 GB
Too big Get file →
Q3_K_S29.9 GB 31.8 GB
Too big Get file →
IQ3_M30.9 GB 32.8 GB
Too big Get file →
Q3_K_M33.3 GB 35.1 GB
Too big Get file →
Q3_K_L36.1 GB 38.0 GB
Too big Get file →
IQ4_XS36.8 GB 38.7 GB
Too big Get file →
Q4_039.0 GB 40.9 GB
Too big Get file →
Q4_K_S39.2 GB 41.1 GB
Too big Get file →
Q4_K_M41.4 GB 43.3 GB
Too big Get file →
Q5_K_S47.5 GB 49.3 GB
Too big Get file →
Q5_K_M48.8 GB 50.6 GB
Too big Get file →

For 8‑GB VRAM, use IQ1_S (14.54 GB file, 17.7 GB memory at 8k context). For 16‑GB VRAM, IQ2_S (21.35 GB file, 24.5 GB memory) or Q2_K (25.46 GB file, 28.6 GB memory). For 24‑GB VRAM, Q3_K_M (33.27 GB file, 36.5 GB memory) or Q4_K_M (41.42 GB file, 44.6 GB memory). Memory estimates are approximate.

What GPU do you need for Llama 2 70b Hf?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: the best quant that fits is IQ2_XS (20.3 GB file, about 23.5 GB in use).
  • 32 GB VRAM: the best quant that fits is IQ3_XS (28.3 GB file, about 31.5 GB in use).
  • 48 GB VRAM: the best quant that fits is Q4_K_M (41.4 GB file, about 44.6 GB in use).

Which Llama 2 70b Hf quant should you choose?

Q4_K_M is the usual starting point for Llama 2 70b Hf: it balances file size and quality well. Here it is a 41.4 GB file.

If memory is tight, the smallest file is IQ1_S at 14.5 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters69B
Layers80
Trained context4,096 tokens
KV cache (fp16)320 KB per token

Source

Quantized bymradermacher
FamilyLlama
Downloads (HF)736
Last updated2024-09-10

Frequently asked questions about Llama 2 70b Hf

How much VRAM does Llama 2 70b Hf need?

About 44.6 GB at Q4_K_M with an 8K context window. That is the 41.4 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Llama 2 70b Hf on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 17.7 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Llama 2 70b Hf GGUF?

It depends on the quant: IQ1_S is 14.5 GB, IQ1_M is 15.9 GB, IQ2_XXS is 18.3 GB, IQ2_XS is 20.3 GB, IQ2_S is 21.4 GB, IQ2_M is 23.2 GB and 14 more.

What context length does Llama 2 70b Hf support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 2.7 GB at 8K and 1.3 GB at 4K.

Where can I download Llama 2 70b Hf GGUF?

The files are on Hugging Face in the mradermacher/Llama-2-70b-hf-i1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Can I run this model on 8 GB VRAM?

Use the IQ1_S quant, which requires about 17.7 GB of VRAM at 8k context.

Does this model support vision tasks?

No, it only processes text input and output.

What license applies to this model?

The model is governed by Meta’s custom commercial license; acceptance is required before download.

Other GGUFs of Llama-2-70b-hf

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Llama models

All Llama →

Other large models

All large models →