Llama 1.24B configuration
Captured config identifies LlamaForCausalLM and llama, with Safetensors metadata reporting 1235814400 parameters.
Open Source Model Profile · ReplaceHumanWithAI
F16_VLLM2
F16_VLLM2 is a 1.24B-parameter Llama text-generation fine-tune from ReplaceHumanWithAI. Its model card identifies unsloth/llama-3.2-1b-instruct-bnb-4bit as the source with Unsloth and TRL training.
- Publisher
- ReplaceHumanWithAI
- Task
- text-generation
- Model type
- llama
- License
- apache-2.0
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
F16_VLLM2 is published by ReplaceHumanWithAI as a llama-based text-generation fine-tune. The captured configuration identifies LlamaForCausalLM and Safetensors metadata reports 1235814400 parameters. According to the model card, it was finetuned from unsloth/llama-3.2-1b-instruct-bnb-4bit with Unsloth and TRL.
Recorded capabilities
Documented Llama 3.2 source
According to the model card, the model was finetuned from unsloth/llama-3.2-1b-instruct-bnb-4bit and developed by ReplaceHumanWithAI.
Unsloth and TRL stack
According to the model card, training used Unsloth with Hugging Face TRL, described as 2x faster.
Apache-2.0 conversational record
Card data records apache-2.0 and hub tags include conversational, sft, unsloth, trl, and text-generation-inference.
Use cases in the source record
- Conversational text generation using the captured transformers Llama configuration with inference-endpoint compatibility.
- Unsloth and TRL fine-tune research building on the documented Llama 3.2 1B instruct source.
Limitations and unknowns
- No evaluation results were extracted from this record.
- No context-window value was extracted from this record.
- Provider state is historical snapshot data, not independently refreshed current availability.
- Source-model and training-speed claims come from a brief publisher model card and were not independently verified by Ethen.
Source and provenance
Source: ReplaceHumanWithAI/F16_VLLM2
Captured: Unknown. Processed: 2026-09-07T19:34:36.265708+00:00.
Uploaded model Developed by: ReplaceHumanWithAI License: apache-2.0 Finetuned from model : unsloth/llama-3.2-1b-instruct-bnb-4bit This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.
F001F002F003F004F005F006F007F008F009F010F011