llm2vec4cxr / llm2vec_config.json
lukeingawesome's picture
Update model with corrected safetensors metadata and proper latent attention loading
2832df4 verified
raw
history blame contribute delete
117 Bytes
{
"pooling_mode": "latent_attention",
"max_length": 512,
"doc_max_length": null,
"skip_instruction": false
}