Granite 20B Code Base R1.1 by ibm-granite

 ยป  All LLMs  ยป  ibm-granite  ยป  Granite 20B Code Base R1.1   URL Share it on

  Arxiv:2405.04324   Autotrain compatible   Code   Codegen   Dataset:bigcode/starcoderdata Dataset:codeparrot/github-code...   Dataset:math-ai/stackmathqa Dataset:open-web-math/open-web...   Endpoints compatible   Gpt bigcode   Granite   Model-index   Region:us   Safetensors   Sharded   Tensorflow

Granite 20B Code Base R1.1 Benchmarks

nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Granite 20B Code Base R1.1 (ibm-granite/granite-20b-code-base-r1.1)

Granite 20B Code Base R1.1 Parameters and Internals

Model Type 
text generation, code generative
Use Cases 
Primary Use Cases:
Code generation, Code explanation, Code fixing, Generating unit tests, Generating documentation, Addressing technical debt issues, Vulnerability detection, Code translation
Limitations:
Generated code is not guaranteed to work as intended, Model generates problematic outputs, Risk of verbatim copying due to memorization in smaller models
Considerations:
Caution urged against complete reliance, potential for problematic outputs and hallucination
Supported Languages 
languages_supported (116 programming languages), proficiency (comprehensive understanding)
Training Details 
Data Sources:
Publicly available datasets (e.g., GitHub Code Clean, Starcoder data), Additional public code repositories and issues from GitHub
Data Volume:
Phase 1: 3 trillion tokens, Phase 2: 1 trillion tokens
Methodology:
Two-phase training strategy
Hardware Used:
IBM's Vela and Blue Vela super computing clusters, NVIDIA A100 and H100 GPUs
Model Architecture:
Decoder-only architecture designed for code-generative tasks
Responsible Ai Considerations 
Mitigation Strategies:
HAP content filter, PII redaction, malware scanning
LLM NameGranite 20B Code Base R1.1
Repository ๐Ÿค—https://huggingface.co/ibm-granite/granite-20b-code-base-r1.1 
Model Size20b
Required VRAM40 GB
Updated2024-12-21
Maintaineribm-granite
Model Typegpt_bigcode
Model Files  5.0 GB: 1-of-9   4.9 GB: 2-of-9   4.9 GB: 3-of-9   4.9 GB: 4-of-9   4.9 GB: 5-of-9   4.9 GB: 6-of-9   4.9 GB: 7-of-9   4.9 GB: 8-of-9   0.7 GB: 9-of-9
Generates CodeYes
Model ArchitectureGPTBigCodeForCausalLM
Licenseapache-2.0
Model Max Length8192
Transformers Version4.41.2
Tokenizer ClassGPT2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size49152
Torch Data Typebfloat16
Activation Functiongelu

Best Alternatives to Granite 20B Code Base R1.1

Best Alternatives
Context / RAM
Downloads
Likes
Granite 20B Code Instruct0K / 40 GB1020930
Granite 20B Functioncalling0K / 40 GB63527
Granite 20B Code Base0K / 40 GB225012
Granite 20B Code Instruct 8K0K / 40 GB98839
Granite 20B Code Base 8K0K / 40 GB164513
Granite 20B Code Instruct R1.10K / 40 GB801
Granite 20B Code Base FP80K / 20.4 GB140
Granite 20B Code Base GGUF0K / 12.8 GB80
Note: green Score (e.g. "73.2") means that the model is better than ibm-granite/granite-20b-code-base-r1.1.

Rank the Granite 20B Code Base R1.1 Capabilities

๐Ÿ†˜ Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! ๐ŸŒŸ

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  

What open-source LLMs or SLMs are you in search of? 40013 in total.

Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241217