dacorvo's picture
dacorvo HF staff
Added Llama-70b batch_size 4 to inference cache
593822e verified
{
"meta-llama/Llama-2-7b-chat-hf": [
{
"batch_size": 1,
"sequence_length": 4096,
"num_cores": 2,
"auto_cast_type": "fp16"
},
{
"batch_size": 1,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 1,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
},
{
"batch_size": 4,
"sequence_length": 4096,
"num_cores": 2,
"auto_cast_type": "fp16"
},
{
"batch_size": 4,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 4,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
},
{
"batch_size": 8,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 8,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
},
{
"batch_size": 16,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 16,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
}
],
"meta-llama/Llama-2-13b-chat-hf": [
{
"batch_size": 1,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 1,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
},
{
"batch_size": 4,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 4,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
},
{
"batch_size": 8,
"sequence_length": 4096,
"num_cores": 8,
"auto_cast_type": "fp16"
},
{
"batch_size": 8,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
}
],
"meta-llama/Llama-2-70b-chat-hf": [
{
"batch_size": 1,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
},
{
"batch_size": 4,
"sequence_length": 4096,
"num_cores": 24,
"auto_cast_type": "fp16"
}
]
}