DE

DeepScaleR 1.5B Preview Abliterated

DeepscalerGGUF by ThomasBaruzierUpdated 2025-02-18
Download on Hugging Face
Parameters1.78B
Context128K
Layers28
Architectureqwen2
Smallest quant0.5 GB
Quick answer

DeepScaleR 1.5B Preview Abliterated needs about 1.9 GB of memory at Q4_K_M (a 1.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 29 quants, from 0.5 GB (IQ1_S) to 3.6 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 0.9 GB.

DeepScaleR 1.5B Preview Abliterated – 1.78B parameter Qwen2 model

DeepScaleR 1.5B Preview Abliterated is a 1.78B parameter Qwen2 architecture model trained on 131072 tokens. It is released under the MIT license and available in multiple GGUF quantizations. The model focuses on efficient inference with a large context window.

Good at

  • Long context handling
  • Efficient inference
  • Compact model size

Limits

  • No vision support
  • Memory estimates based on 8k context
  • MIT license allows modification

Best for

  • Chatbot prototypes
  • Research experiments
  • Edge inference

Who it’s for. Developers building lightweight LLMs and researchers needing a 1.5B Qwen2 variant for experimentation or deployment on modest hardware.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S0.5 GB 1.2 GB
Fits Get file →
IQ1_M0.5 GB 1.3 GB
Fits Get file →
IQ2_XXS0.6 GB 1.3 GB
Fits Get file →
IQ2_XS0.6 GB 1.4 GB
Fits Get file →
IQ2_S0.7 GB 1.4 GB
Fits Get file →
IQ2_M0.7 GB 1.4 GB
Fits Get file →
Q2_K_S0.7 GB 1.5 GB
Fits Get file →
Q2_K0.8 GB 1.5 GB
Fits Get file →
IQ3_XXS0.8 GB 1.5 GB
Fits Get file →
IQ3_XS0.8 GB 1.6 GB
Fits Get file →
Q3_K_S0.9 GB 1.6 GB
Fits Get file →
IQ3_S0.9 GB 1.6 GB
Fits Get file →
IQ3_M0.9 GB 1.6 GB
Fits Get file →
Q3_K_M0.9 GB 1.7 GB
Fits Get file →
Q3_K_L1.0 GB 1.7 GB
Fits Get file →
Q4_01.0 GB 1.7 GB
Fits Get file →
IQ4_XS1.0 GB 1.8 GB
Fits Get file →
IQ4_NL1.1 GB 1.8 GB
Fits Get file →
Q4_01.1 GB 1.8 GB
Fits Get file →
Q4_K_S1.1 GB 1.8 GB
Fits Get file →
Q4_K_M1.1 GB 1.9 GB
Fits Get file →
Q4_11.2 GB 1.9 GB
Fits Get file →
Q5_K_S1.3 GB 2.0 GB
Fits Get file →
Q5_01.3 GB 2.0 GB
Fits Get file →
Q5_K_M1.3 GB 2.0 GB
Fits Get file →
Q5_11.4 GB 2.1 GB
Fits Get file →
Q6_K1.5 GB 2.2 GB
Fits Get file →
Q8_01.9 GB 2.6 GB
Fits Get file →
F163.6 GB 4.3 GB
Fits Get file →

For 8GB VRAM, use IQ1_S (0.51 GB) or IQ1_M (0.54 GB). For 12GB VRAM, Q3_K_M (0.92 GB) or Q4_K_M (1.12 GB) are suitable. For 16GB VRAM, Q5_K_S (1.26 GB) or Q6_K (1.46 GB) provide higher precision.

What GPU do you need for DeepScaleR 1.5B Preview Abliterated?

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 (1.9 GB file, about 2.6 GB in use).
  • 6 GB VRAM: the best quant that fits is F16 (3.6 GB file, about 4.3 GB in use).

Which DeepScaleR 1.5B Preview Abliterated quant should you choose?

Q4_K_M is the usual starting point for DeepScaleR 1.5B Preview Abliterated: it balances file size and quality well. Here it is a 1.1 GB file.

The Q8_0 file is 1.9 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 IQ1_S at 0.5 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

Architectureqwen2
Parameters1.78B
Layers28
Trained context131,072 tokens
KV cache (fp16)28 KB per token

Source

Quantized byThomasBaruzier
Licensemit
Downloads (HF)1,360
Last updated2025-02-18

Frequently asked questions about DeepScaleR 1.5B Preview Abliterated

How much VRAM does DeepScaleR 1.5B Preview Abliterated need?

About 1.9 GB at Q4_K_M with an 8K context window. That is the 1.1 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 DeepScaleR 1.5B Preview Abliterated on an 8 GB GPU?

Yes, with F16: the 3.6 GB file needs about 4.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of DeepScaleR 1.5B Preview Abliterated GGUF?

It depends on the quant: IQ1_S is 0.5 GB, IQ1_M is 0.5 GB, IQ2_XXS is 0.6 GB, IQ2_XS is 0.6 GB, IQ2_S is 0.7 GB, IQ2_M is 0.7 GB and 23 more.

What context length does DeepScaleR 1.5B Preview Abliterated support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 0.2 GB at 8K and 0.9 GB at 32K.

Where can I download DeepScaleR 1.5B Preview Abliterated GGUF?

The files are on Hugging Face in the ThomasBaruzier/DeepScaleR-1.5B-Preview-abliterated-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does DeepScaleR 1.5B Preview Abliterated use?

Hugging Face lists the license as mit. 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 IQ1_S on 4GB VRAM?

IQ1_S requires about 1.2 GB of VRAM at 8k context, so 4GB VRAM is sufficient for inference.

What is the maximum context length?

The model was trained on 131072 tokens, but memory estimates are provided for 8k context.

Is the model licensed under MIT?

Yes, the model is released under the MIT license, allowing free use and modification.

Similar models

Browse all →

More from ThomasBaruzier

All by ThomasBaruzier →

More Deepscaler models

All Deepscaler →

Other tiny models

All tiny models →