PrunaAI

/

sapienzanlp-modello-italia-9b-bf16-QUANTO-int8bit-smashed

like 0

Follow

Pruna AI 137

• "torch.int8",
• "transformers.generation.configuration_utils.GenerationConfig",
• "quanto.tensor.qtype.qtype",
• "torch._utils._rebuild_parameter",
• "torch.nn.modules.container.ModuleList",
• "transformers.activations.FastGELUActivation",
• "collections.OrderedDict",
• "torch.nn.modules.sparse.Embedding",
• "__builtin__.set",
• "quanto.nn.qlinear.QLinear",
• "torch._utils._rebuild_tensor_v2",
• "transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXForCausalLM",
• "torch.nn.modules.dropout.Dropout",
• "torch.BoolStorage",
• "transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXLayer",
• "torch.BFloat16Storage",
• "transformers.models.gpt_neox.configuration_gpt_neox.GPTNeoXConfig",
• "torch.device",
• "transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXModel",
• "torch.FloatStorage",
• "transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXSdpaAttention",
• "torch.nn.modules.normalization.LayerNorm",
• "torch.bfloat16",
• "transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXRotaryEmbedding",
• "transformers.models.gpt_neox.modeling_gpt_neox.GPTNeoXMLP"

How to fix it?