LLM EXPLORER 61,491 MODELS INDEXED

Falcon 7B Instruct Sharded by vilsonrodrigues

By vilsonrodrigues · 10197 downloads
HF Repo ↗ Copy URL

Falcon 7B Instruct Sharded is an open-source language model by vilsonrodrigues. Features: 7b LLM, VRAM: 27.7GB, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.1.

  Arxiv:1911.02150   Arxiv:2005.14165   Arxiv:2104.09864   Arxiv:2205.14135   Arxiv:2306.01116   Custom code Dataset:tiiuae/falcon-refinedw...   En   Endpoints compatible   Falcon   Instruct   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/vilsonrodrigues/falcon-7b-instruct-sharded 

Falcon 7B Instruct Sharded Parameters and Internals

Model Type 
Causal decoder-only
Use Cases 
Areas:
Research, Chat applications
Applications:
Instruct models, Chatbots
Primary Use Cases:
Chat/instruct tasks
Limitations:
Trained mostly on English data, may not generalize well to other languages
Considerations:
Guardrails recommended for production use
Additional Notes 
Not optimized for further finetuning; intended for instruct/chat scenarios.
Supported Languages 
English (native proficiency), French (native proficiency)
Training Details 
Data Sources:
Bai ze, GPT4All, GPTeacher, RefinedWeb-English
Data Volume:
250M tokens
Methodology:
Finetuned on a mixture of instruct/chat datasets
Context Length:
2048
Hardware Used:
32 A100 40GB GPUs
Model Architecture:
Architecture adapted from GPT-3, with rotary positional embeddings, multiquery and FlashAttention
Input Output 
Input Format:
Text prompts
Accepted Modalities:
Text
Output Format:
Generated text
Performance Tips:
Use PyTorch 2.0 and appropriate hardware for fastest inference
LLM NameFalcon 7B Instruct Sharded
Repository 🤗https://huggingface.co/vilsonrodrigues/falcon-7b-instruct-sharded 
Model Size7b
Required VRAM27.7 GB
Updated2026-07-18
Maintainervilsonrodrigues
Model Typefalcon
Instruction-BasedYes
Model Files  1.7 GB: 1-of-15   2.0 GB: 2-of-15   1.8 GB: 3-of-15   2.0 GB: 4-of-15   2.0 GB: 5-of-15   1.8 GB: 6-of-15   2.0 GB: 7-of-15   2.0 GB: 8-of-15   1.8 GB: 9-of-15   2.0 GB: 10-of-15   2.0 GB: 11-of-15   1.8 GB: 12-of-15   2.0 GB: 13-of-15   2.0 GB: 14-of-15   0.8 GB: 15-of-15
Supported Languagesen
Model ArchitectureFalconForCausalLM
Licenseapache-2.0
Model Max Length2048
Transformers Version4.27.4
Is Biased0
Tokenizer ClassPreTrainedTokenizerFast
Vocabulary Size65024
Torch Data Typebfloat16

Best Alternatives to Falcon 7B Instruct Sharded

Best Alternatives
Context / RAM
Downloads
Likes
Claire 7B FR Instruct 0.12K / 13.9 GB116
... Openassistant Toxicity Reduce2K / 30.4 GB80
New Falcon2K / 4.3 GB51
Claire 7B 0.1 Instruct2K / 7.2 GB31
New Falcon2K / 4.3 GB00
Falcon 7B Instruct0K / 14.4 GB292711032
Falcon 7B Instruct Quant0K / 4 GB11
Falcon 7B Instruct Ft0K / 13.9 GB80
Falcon 7B IT Technical Support0K / 13.8 GB141
Falcon Tiny Random0K / 0 GB1700
Note: green Score (e.g. "73.2") means that the model is better than vilsonrodrigues/falcon-7b-instruct-sharded.