Skip to content

EthenEthenEthen

LoginPlatformTry Ethen

Open Source Model Profile · unsloth

Qwen2.5-72B-Instruct

Qwen2.5-72B-Instruct from unsloth is a 72.71B-parameter Qwen2 instruction-tuned model. Its card documents 80 layers, grouped-query attention, and 131,072-token context handling.

Publisher
unsloth
Task
text-generation
Model type
qwen2
License
other
Library
transformers
Publication status
Accepted · not indexed

Model overview

Qwen2.5-72B-Instruct from unsloth is published as a Qwen2-based instruction-tuned text-generation model. The captured configuration identifies Qwen2ForCausalLM and Safetensors metadata reports 72706203648 parameters. Hub tags cite Qwen/Qwen2.5-72B-Instruct as its base, and the card describes the instruction-tuned 72B Qwen2.5 model with 80 layers and a 131,072-token context.

Recorded capabilities

72.71B Qwen2 scale

Captured metadata reports 72706203648 parameters for this Qwen2ForCausalLM record, and the card states 72.7B total with 70.0B non-embedding.

131K documented context

According to the model card, the model handles a full 131,072-token context with 8192-token generation.

Documented 80-layer GQA design

According to the model card, the model has 80 layers with 64 Q and 8 KV grouped-query attention heads, using RoPE, SwiGLU, RMSNorm, and QKV bias.

Qwen2.5-72B-Instruct lineage

Hub tags cite Qwen/Qwen2.5-72B-Instruct as base and fine-tune source, and the card describes the instruction-tuned 72B Qwen2.5 model.

Publisher finetuning notebooks

According to the model card, Unsloth provides free beginner-friendly Colab notebooks exportable to GGUF, vLLM, or Hugging Face.

Use cases in the source record

  • Long-context instruction workflows using the documented 131,072-token context handling and chat-template tokenization.
  • Multilingual conversational experiments across the English, Chinese, French, Spanish, Portuguese, German, Italian, Russian, Japanese, Korean, Vietnamese, Thai, and Arabic tags on the record.

Limitations and unknowns

  • No scored evaluation results were extracted from this record.
  • Provider state is historical snapshot data, not independently refreshed current availability.
  • The captured license value is 'other' with no specific license named in the extracted card claims, so usage rights need direct confirmation.
  • Speed and memory claims for Unsloth finetuning are publisher statements that Ethen has not independently verified.

Source and provenance

Source: unsloth/Qwen2.5-72B-Instruct

Captured: Unknown. Processed: 2026-09-07T19:35:00.313504+00:00.

Finetune Llama 3.1, Gemma 2, Mistral 2-5x faster with 70% less memory via Unsloth! We have a Qwen 2.5 (all model sizes) free Google Colab Tesla T4 notebook . Also a Qwen 2.5 conversational style notebook . ✨ Finetune for Free All notebooks are beginner friendly ! Add your dataset, click "Run All", and you'll get a 2x faster finetuned model which can be exported to GGUF, vLLM or uploaded to Hugging Face. Unsloth supports Free Notebooks Performance Memory use Llama-3.1 8b ▶️ Start on Colab 2.4x faster 58% less Phi-3.5 (mini) ▶️ Start on Colab 2x faster 50% less Gemma-2 9b ▶️ Start on Colab 2.4x faster 58% less Mistral 7b ▶️ Start on C…

F001F002F003F004F005F006F007F009F010F011F012F013F014F015F016F017F018F019