Closed weigary closed 2 months ago
Hello,
Thank you for opening. We ll take a look
We are not using the latest version and updating it. Here is the config for ShieldGemma
{
"model_id": "google/shieldgemma-27b",
"instance": "a2-ultragpu-1g",
"configuration": {
"model_id": "google/shieldgemma-27b",
"max_batch_prefill_tokens": 8192,
"max_input_length": 6000,
"max_total_tokens": 6144,
"num_shard": 1,
"quantize": null,
"estimated_memory_in_gigabytes": 66.96
}
}
Hi,
Now we are making http request to this URL
https://huggingface.co/api/integrations/tgi/v1/provider/gcp/recommend?model_id=google/shieldgemma-27b
to get the recommended deployment configs. But we found it returns{"error":"No recommendation found"}
for a number of new models. Can we refresh the model list?Thanks!