32.76B Qwen2 function-calling record
Captured Safetensors metadata reports 32,763,876,352 parameters for a Qwen2ForCausalLM record with model type qwen2.
Open Source Model Profile · Salesforce
xLAM-2-32b-fc-r
xLAM-2-32b-fc-r is a 32.76B-parameter Salesforce Qwen2 text-generation fine-tune for multi-turn conversation and function calling. Its card documents Hugging Face chat-template and vLLM inference support.
- Publisher
- Salesforce
- Task
- text-generation
- Model type
- qwen2
- License
- cc-by-nc-4.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
xLAM-2-32b-fc-r is published by Salesforce as a text-generation model in the xLAM-2 family. The captured configuration identifies Qwen2ForCausalLM with model type qwen2, and Safetensors metadata reports 32,763,876,352 parameters. The recorded license is cc-by-nc-4.0, and hub tags mark function-calling, tool-use, conversational, and the APIGen-MT-5k and xlam-function-calling-60k datasets.
Recorded capabilities
APIGen-MT training lineage
According to the model card, the series is trained with the APIGen-MT framework on simulated agent-human interactions, and hub tags name the 60k function-calling and 5k APIGen-MT datasets.
Chat template and vLLM support
The card documents Hugging Face chat-template use and states compatibility with vLLM and Transformers-based inference, with vLLM 0.6.5 or later recommended for Qwen-based models.
Use cases in the source record
- Multi-turn conversational tool-use workflows using the card’s documented Hugging Face chat template with supplied tools.
- Function-calling inference experiments served with vLLM, which the card describes as compatible with all models in the series.
Limitations and unknowns
- No numeric evaluation scores were extracted; the card describes BFCL v3 and τ-bench comparisons without captured scores, so publisher performance language remains unverified.
- The card lists 32k default context with up to 128k via YaRN techniques for Qwen-2.5-based models; this is a publisher claim and was not independently verified.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
Source and provenance
Source: Salesforce/xLAM-2-32b-fc-r
Captured: Unknown. Processed: 2026-09-07T19:35:15.476029+00:00.
[Paper] | [Homepage] | [Dataset] | [Github] Welcome to the xLAM-2 Model Family! Large Action Models (LAMs) are advanced language models designed to enhance decision-making by translating user intentions into executable actions. As the brains of AI agents , LAMs autonomously plan and execute tasks to achieve specific goals, making them invaluable for automating workflows across diverse domains. This model release is for research purposes only. The new xLAM-2 series, built on our most advanced data synthesis, processing, and training pipelines, marks a significant leap in multi-turn conversation and tool usage . Trained using our nove…
F001F002F003F004F005F006F007F008F009F010F013F014F017F018F019F020F021F022F024F029