convaiinnovations/laya
Text Classification • 0.4B • Updated • 47.8k • • 5.61k
Enterprise Fleet
Models staged for Aaron's enterprise fleet — saved on the Jarvis proxy page for later evaluation and use.
Note System 1 decision model (RLCD) — Ferrum/Laya candidate for later evaluation.
convaiinnovations/laya-multilingual
Text Classification • 0.3B • Updated • 11k • • 390Note Multilingual Laya variant — saved with the fleet for later evaluation.
MoritzLaurer/deberta-v3-large-zeroshot-v2.0
Zero-Shot Classification • 0.4B • Updated • 150k • • 145Note DECISION | 435M zero-shot classification / routing, MIT
sentence-transformers/all-MiniLM-L6-v2
Sentence Similarity • 22.7M • Updated • 218M • • 6.34kNote EMBED | 23M tiny embeddings, high-throughput tagging/search
BAAI/bge-small-en-v1.5
Feature Extraction • 33.4M • Updated • 59.4M • • 643Note EMBED | 33M small retrieval embeddings, MIT
BAAI/bge-base-en-v1.5
Feature Extraction • 0.1B • Updated • 9.23M • • 532Note EMBED | 109M mid retrieval embeddings, MIT
BAAI/bge-large-en-v1.5
Feature Extraction • 0.3B • Updated • 10.1M • • 773Note EMBED | 335M quality retrieval embeddings, MIT
BAAI/bge-reranker-base
Text Classification • 0.3B • Updated • 3.75M • • 258Note RERANK | 278M reranker for RAG quality, MIT
BAAI/bge-reranker-large
Feature Extraction • 0.6B • Updated • 2.21M • • 480Note RERANK | 560M reranker, higher quality, MIT
Qwen/Qwen2.5-0.5B-Instruct
Text Generation • 0.5B • Updated • 6.28M • • 689Note EDGE LLM | 0.5B endpoint/embedded triage
Qwen/Qwen2.5-1.5B-Instruct
Text Generation • 2B • Updated • 6.02M • • 887Note EDGE LLM | 1.5B light local assistant
HuggingFaceTB/SmolLM2-1.7B-Instruct
Text Generation • 2B • Updated • 227k • 763Note EDGE LLM | 1.7B efficient small instruct
Qwen/Qwen2.5-3B-Instruct
Text Generation • 3B • Updated • 4.3M • • 603Note EDGE LLM | 3B capable small local model
microsoft/Phi-3.5-mini-instruct
Text Generation • 4B • Updated • 359k • 1.28kNote EDGE LLM | 3.8B small reasoning-leaning, MIT
mistralai/Mistral-7B-Instruct-v0.3
7B • Updated • 2.04M • 3.8kNote WORKHORSE | 7B general instruct, Apache-2.0
Qwen/Qwen2.5-7B-Instruct
Text Generation • 8B • Updated • 7.36M • • 2.44kNote WORKHORSE | 7B general/RAG default
Qwen/Qwen2.5-14B-Instruct
Text Generation • 15B • Updated • 1.12M • • 375Note WORKHORSE+ | 14B deeper analysis, local server
deepseek-ai/DeepSeek-R1-Distill-Qwen-14B
Text Generation • 15B • Updated • 330k • • 719Note REASONING | 14B distilled reasoning, MIT
Qwen/Qwen2.5-32B-Instruct
Text Generation • 33B • Updated • 1.61M • • 370Note HEAVY | 32B high-capability server model
deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
Text Generation • 33B • Updated • 379k • • 1.66kNote REASONING | 32B distilled reasoning, MIT
Qwen/Qwen2.5-72B-Instruct
Text Generation • 73B • Updated • 277k • • 1.02kNote HEAVY | 72B ceiling reference model
Qwen/Qwen2.5-Coder-1.5B-Instruct
Text Generation • 2B • Updated • 464k • • 159Note CODE | 1.5B completion/light edits
Qwen/Qwen2.5-Coder-7B-Instruct
Text Generation • 8B • Updated • 1.59M • • 831Note CODE | 7B everyday coding assistant
Qwen/Qwen2.5-Coder-14B-Instruct
Text Generation • 15B • Updated • 1.76M • • 201Note CODE | 14B repo-level coding
Qwen/Qwen2.5-Coder-32B-Instruct
Text Generation • 33B • Updated • 775k • • 2.18kNote CODE | 32B heavy code review/generation
Qwen/Qwen2.5-VL-3B-Instruct
Image-Text-to-Text • 4B • Updated • 2.02M • • 726Note VISION | 3B screenshots/docs, light
Qwen/Qwen2.5-VL-7B-Instruct
Image-Text-to-Text • 8B • Updated • 5.13M • • 1.75kNote VISION | 7B screenshots/docs, default
Qwen/Qwen2.5-VL-32B-Instruct
Image-Text-to-Text • 33B • Updated • 770k • • 505Note VISION | 32B hard documents/visual analysis
openai/whisper-small
Automatic Speech Recognition • 0.2B • Updated • 3.33M • 628Note AUDIO | 242M fast transcription
openai/whisper-large-v3-turbo
Automatic Speech Recognition • 0.8B • Updated • 6.34M • • 3.45kNote AUDIO | 809M quality transcription, MIT
microsoft/layoutlmv3-base
0.1B • Updated • 926k • 521Note DOC | 125M document layout — WARNING: CC-BY-NC-SA non-commercial license, eval only