Nova Jais 2 70b v2 needs about 61.6 GB of memory at Q4_K_M (a 45.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 24 quants, from 0.0 GB to 59.1 GB (Q6_K).
Nova Jais 2 70b v2 is a 72B-parameter Nova model in GGUF format, quantized by mradermacher. Its trained context length is 8K tokens. The repository has 24 quant files, from 0.0 GB to 59.1 GB (Q6_K).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 16.5 GB | Too big | Get file → | |
| IQ1_S | 15.9 GB | 32.4 GB | Too big | Get file → | |
| IQ1_M | 17.3 GB | 33.8 GB | Too big | Get file → | |
| IQ2_XXS | 19.6 GB | 36.1 GB | Too big | Get file → | |
| IQ2_XS | 21.6 GB | 38.1 GB | Too big | Get file → | |
| IQ2_S | 22.9 GB | 39.4 GB | Too big | Get file → | |
| IQ2_M | 24.7 GB | 41.2 GB | Too big | Get file → | |
| Q2_K_S | 25.0 GB | 41.4 GB | Too big | Get file → | |
| IQ3_XXS | 27.7 GB | 44.2 GB | Too big | Get file → | |
| Q2_K | 28.7 GB | 45.1 GB | Too big | Get file → | |
| IQ3_XS | 30.1 GB | 46.6 GB | Too big | Get file → | |
| IQ3_S | 31.4 GB | 47.9 GB | Too big | Get file → | |
| Q3_K_S | 32.3 GB | 48.8 GB | Too big | Get file → | |
| IQ3_M | 33.2 GB | 49.7 GB | Too big | Get file → | |
| Q3_K_M | 36.7 GB | 53.2 GB | Too big | Get file → | |
| IQ4_XS | 38.6 GB | 55.1 GB | Too big | Get file → | |
| Q3_K_L | 40.4 GB | 56.9 GB | Too big | Get file → | |
| Q4_0 | 41.0 GB | 57.5 GB | Too big | Get file → | |
| Q4_K_S | 41.7 GB | 58.2 GB | Too big | Get file → | |
| Q4_K_M | 45.1 GB | 61.6 GB | Too big | Get file → | |
| Q4_1 | 45.3 GB | 61.7 GB | Too big | Get file → | |
| Q5_K_S | 49.7 GB | 66.2 GB | Too big | Get file → | |
| Q5_K_M | 51.8 GB | 68.3 GB | Too big | Get file → | |
| Q6_K | 59.1 GB | 75.6 GB | Too big | Get file → |
What GPU do you need for Nova Jais 2 70b v2?
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 the main file (0.0 GB file, about 16.5 GB in use).
- 32 GB VRAM: the best quant that fits is the main file (0.0 GB file, about 16.5 GB in use).
- 48 GB VRAM: the best quant that fits is IQ3_S (31.4 GB file, about 47.9 GB in use).
Which Nova Jais 2 70b v2 quant should you choose?
Q4_K_M is the usual starting point for Nova Jais 2 70b v2: it balances file size and quality well. Here it is a 45.1 GB file.
If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Nova Jais 2 70b v2
How much VRAM does Nova Jais 2 70b v2 need?
About 61.6 GB at Q4_K_M with an 8K context window. That is the 45.1 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 Nova Jais 2 70b v2 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 16.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Nova Jais 2 70b v2 GGUF?
It depends on the quant: file is 0.0 GB, IQ1_S is 15.9 GB, IQ1_M is 17.3 GB, IQ2_XXS is 19.6 GB, IQ2_XS is 21.6 GB, IQ2_S is 22.9 GB and 18 more.
What context length does Nova Jais 2 70b v2 support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 16.0 GB at 8K and 16.0 GB at 8K.
Where can I download Nova Jais 2 70b v2 GGUF?
The files are on Hugging Face in the mradermacher/nova-jais-2-70b-v2-i1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.