Llama3Ternary 2b needs about 2.1 GB of memory at Q4_K_M (a 1.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 5 quants, from 0.7 GB (TQ1_0) to 4.2 GB (F16).
Llama3Ternary 2b – 2.1B ternary‑quantized Llama model with tool‑calling
Llama3Ternary 2b is a 2.1B parameter decoder‑only transformer natively trained with ternary weights (1.58‑bit). It supports text generation and built‑in tool calling. The model is released under the bsl‑1.0 license.
Good at
- Tool‑calling
- Text generation
- Ternary quantization
Limits
- No vision support
- Limited to 8192 context tokens
- Requires written agreement for access
Best for
- Embedded LLMs with low VRAM
- Tool‑enabled chatbots
- Edge inference
Who it’s for. Developers needing a small, tool‑aware Llama model for edge or embedded deployments. The model is suitable for applications that can operate within the 8192 token context limit and require a lightweight quantization scheme.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| TQ1_0 | 0.7 GB | 1.5 GB | Fits | Get file → | |
| TQ2_0 | 0.8 GB | 1.5 GB | Fits | Get file → | |
| Q4_K_M | 1.3 GB | 2.1 GB | Fits | Get file → | |
| Q8_0 | 2.2 GB | 3.0 GB | Fits | Get file → | |
| F16 | 4.2 GB | 5.0 GB | Fits | Get file → |
For 8 GB VRAM, the TQ1_0 or TQ2_0 quantizations use about 1.5 GB of VRAM at 8 k context. If more memory is available, Q4_K_M uses about 2.1 GB. The Q8_0 and F16 quantizations require 3 GB and 5 GB respectively and are suitable for higher‑precision workloads.
What GPU do you need for Llama3Ternary 2b?
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: the best quant that fits is Q8_0 (2.2 GB file, about 3.0 GB in use).
- 6 GB VRAM: the best quant that fits is F16 (4.2 GB file, about 5.0 GB in use).
Which Llama3Ternary 2b quant should you choose?
Q4_K_M is the usual starting point for Llama3Ternary 2b: it balances file size and quality well. Here it is a 1.3 GB file.
The Q8_0 file is 2.2 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is TQ1_0 at 0.7 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Llama3Ternary 2b
How much VRAM does Llama3Ternary 2b need?
About 2.1 GB at Q4_K_M with an 8K context window. That is the 1.3 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 Llama3Ternary 2b on an 8 GB GPU?
Yes, with F16: the 4.2 GB file needs about 5.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Llama3Ternary 2b GGUF?
It depends on the quant: TQ1_0 is 0.7 GB, TQ2_0 is 0.8 GB, Q4_K_M is 1.3 GB, Q8_0 is 2.2 GB, F16 is 4.2 GB.
What context length does Llama3Ternary 2b support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.3 GB at 8K and 0.3 GB at 8K.
Where can I download Llama3Ternary 2b GGUF?
The files are on Hugging Face in the CMSManhattan/Llama3Ternary_2b repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Llama3Ternary 2b use?
Hugging Face lists the license as bsl-1.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Llama3Ternary 2b on 8 GB VRAM?
Yes, the TQ1_0 and TQ2_0 quantizations use approximately 1.5 GB of VRAM at 8 k context, fitting comfortably on 8 GB.
Does the model support vision tasks?
No, the model has no vision capability.
What license applies to this model?
The model is released under the bsl‑1.0 license.