prithivMLmods/SmolLM2-1.7B-F32-GGUF

SmolLM2-1.7B-Instruct-GGUF

SmolLM2-1.7B-Instruct : The 1.7B variant demonstrates significant advances over its predecessor SmolLM1-1.7B, particularly in instruction following, knowledge, reasoning, and mathematics. It was trained on 11 trillion tokens using a diverse dataset combination: FineWeb-Edu, DCLM, The Stack, along with new mathematics and coding datasets that we curated and will release soon. We developed the instruct version through supervised fine-tuning (SFT) using a combination of public datasets and our own curated datasets. We then applied Direct Preference Optimization (DPO) using UltraFeedback.

Model Files

File Name	Size	Format Description
SmolLM2-1.7B-Instruct.F32.gguf	6.85 GB	Full precision (32-bit floating point)
SmolLM2-1.7B-Instruct.BF16.gguf	3.42 GB	Brain floating point 16-bit
SmolLM2-1.7B-Instruct.F16.gguf	3.42 GB	Half precision (16-bit floating point)
SmolLM2-1.7B-Instruct.Q8_0.gguf	1.82 GB	8-bit quantization
SmolLM2-1.7B-Instruct.Q6_K.gguf	1.41 GB	6-bit quantization (K-quant)
SmolLM2-1.7B-Instruct.Q5_K_M.gguf	1.23 GB	5-bit quantization (K-quant, medium)
SmolLM2-1.7B-Instruct.Q5_K_S.gguf	1.19 GB	5-bit quantization (K-quant, small)
SmolLM2-1.7B-Instruct.Q4_K_M.gguf	1.06 GB	4-bit quantization (K-quant, medium)
SmolLM2-1.7B-Instruct.Q4_K_S.gguf	999 MB	4-bit quantization (K-quant, small)
SmolLM2-1.7B-Instruct.Q3_K_L.gguf	933 MB	3-bit quantization (K-quant, large)
SmolLM2-1.7B-Instruct.Q3_K_M.gguf	860 MB	3-bit quantization (K-quant, medium)
SmolLM2-1.7B-Instruct.Q3_K_S.gguf	777 MB	3-bit quantization (K-quant, small)
SmolLM2-1.7B-Instruct.Q2_K.gguf	675 MB	2-bit quantization (K-quant)

Quants Usage

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):