What are the hardware requirements for Prem 1B Chat 32K?

Prem 1B Chat 32K requires approximately 1.4 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Prem 1B Chat 32K and how large is it?

Prem 1B Chat 32K is developed by ucalyptus, a model with 1b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Prem 1B Chat 32K?

Prem 1B Chat 32K is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Prem 1B Chat 32K by ucalyptus — VRAM 1.4GB

Name: Prem 1B Chat 32K
Author: ucalyptus

Prem 1B Chat 32K is an open-source language model by ucalyptus. Features: 1b LLM, VRAM: 1.4GB, License: apache-2.0, LLM Explorer Score: 0.12.

Arxiv:2309.10400 Adapter Base model:adapter:prem-resear... Base model:prem-research/prem-... Finetuned Generated from trainer Llama Lora Peft Region:us Safetensors Tensorboard

Model Card on HF 🤗: https://huggingface.co/ucalyptus/prem-1B-chat-32k

Prem 1B Chat 32K Parameters and Internals

Model Type

LlamaForCausalLM

Additional Notes

The model was built using Axolotl, version 0.4.0, and employs strategies such as using a learning rate of 1e-05, Adam optimizer, and a cosine learning rate scheduler. Gradient checkpointing is used with multi-device setup.

Training Details

Data Sources:

Argilla DPO dataset

Methodology:

The model is fine-tuned on the Argilla DPO dataset with 32K context extended using PoSE method.

Context Length:

8192

Hardware Used:

multi-GPU, num_devices: 8

LLM Name	Prem 1B Chat 32K
Repository 🤗	https://huggingface.co/ucalyptus/prem-1B-chat-32k
Base Model(s)	Prem 1B Chat premai-io/prem-1B-chat
Model Size	1b
Required VRAM	1.4 GB
Updated	2026-08-05
Maintainer	ucalyptus
Model Files	1.4 GB
Model Architecture	Adapter
License	apache-2.0
Is Biased	none
Tokenizer Class	LlamaTokenizer
Padding Token	<\|end_of_text\|>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	q_proj\|o_proj\|v_proj\|k_proj
LoRA Alpha	1024
LoRA Dropout	0.1
R Param	1024

Rank the Prem 1B Chat 32K capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Prem 1B Chat 32K

Best Alternatives	Context / RAM	Downloads	Likes
...ma 3 1B It Reasoning Grpo Lora	0K / 0.2 GB	14	4
Openhermes 1B Olmo Sft Qlora	0K / 0 GB	7	0
Zephyr 1B Olmo Sft Qlora	0K / 0 GB	5	1
...iabolic6045 ELN AOC CAIN QLoRA	0K / 0 GB	5	0
Llama 3.2 1B Airoboros Merged	0K / 0.1 GB	5	0
...la Alpaca Orca Instruct V0.0.1	0K / 1.1 GB	88	0
Peft Starcoder Lora Python	0K / 0 GB	14	0
Prem 1B 32K	0K / 1.4 GB	7	1
...ama 3 1M Context Gradient Lora	0K / 3.5 GB	0	12
Falcon Arxiv Long Summary 1B	0K / 0.2 GB	4	1

Note: green Score (e.g. "73.2") means that the model is better than ucalyptus/prem-1B-chat-32k.