PrunaAI

/

ISTA-DASLab-Meta-Llama-3-70B-Instruct-AQLM-2Bit-1x16-QUANTO-float8bit-smashed

like 0

Follow

Pruna AI 118

• "transformers.generation.configuration_utils.GenerationConfig",
• "transformers.models.llama.modeling_llama.LlamaMLP",
• "__builtin__.set",
• "torch.nn.modules.activation.SiLU",
• "torch.nn.modules.container.ModuleList",
• "transformers.models.llama.modeling_llama.LlamaRotaryEmbedding",
• "torch.nn.modules.sparse.Embedding",
• "collections.OrderedDict",
• "torch.nn.modules.linear.Linear",
• "torch._utils._rebuild_parameter",
• "transformers.models.llama.modeling_llama.LlamaDecoderLayer",
• "torch.device",
• "transformers.models.llama.modeling_llama.LlamaSdpaAttention",
• "transformers.utils.quantization_config.AqlmConfig",
• "torch.float8_e4m3fn",
• "torch.HalfStorage",
• "transformers.utils.quantization_config.QuantizationMethod",
• "transformers.quantizers.quantizer_aqlm.AqlmHfQuantizer",
• "torch.ShortStorage",
• "torch._utils._rebuild_tensor_v2",
• "transformers.models.llama.modeling_llama.LlamaModel",
• "quanto.tensor.qtype.qtype",
• "aqlm.inference.QuantizedLinear",
• "torch.FloatStorage",
• "transformers.models.llama.modeling_llama.LlamaRMSNorm",
• "transformers.models.llama.modeling_llama.LlamaForCausalLM",
• "transformers.models.llama.configuration_llama.LlamaConfig",
• "torch.float16",
• "quanto.nn.qlinear.QLinear"

How to fix it?