diff --git a/results/inference/Llama2-7B/llama2_predict_1.log b/results/inference/Llama2-7B/llama2_predict_1.log new file mode 100644 index 00000000..d53622a4 --- /dev/null +++ b/results/inference/Llama2-7B/llama2_predict_1.log @@ -0,0 +1,58 @@ +08/18/2024 17:34:57 - INFO - llamafactory.cli - Initializing distributed tasks at: 127.0.0.1:23898 +08/18/2024 17:35:04 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:04 - INFO - llamafactory.hparams.parser - Process rank: 6, device: cuda:6, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:05 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:05 - INFO - llamafactory.hparams.parser - Process rank: 4, device: cuda:4, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:05 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:05 - INFO - llamafactory.hparams.parser - Process rank: 5, device: cuda:5, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:05 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:05 - INFO - llamafactory.hparams.parser - Process rank: 1, device: cuda:1, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:05 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:05 - INFO - llamafactory.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:05 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:05 - INFO - llamafactory.hparams.parser - Process rank: 3, device: cuda:3, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:05 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:35:05 - INFO - llamafactory.hparams.parser - Process rank: 2, device: cuda:2, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:35:06 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:35:16 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:35:16 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:35:16 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:35:16 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:35:16 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:35:16 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +eval example: +input_ids: +[1, 518, 25580, 29962, 29871, 30982, 31695, 31863, 31577, 30210, 30457, 30502, 31302, 30858, 30267, 518, 29914, 25580, 29962] +inputs: + [INST] 保持健康的三个提示。 [/INST] +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:35:34 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:36:13 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:13 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:36:13 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:13 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:36:14 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:14 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:14 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:14 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:14 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:36:14 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:36:14 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:36:14 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:36:14 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:36:14 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +***** predict metrics ***** + predict_bleu-4 = 1.4552 + predict_model_preparation_time = 0.0031 + predict_rouge-1 = 11.0984 + predict_rouge-2 = 3.0446 + predict_rouge-l = 5.0312 + predict_runtime = 0:01:03.50 + predict_samples_per_second = 0.787 + predict_steps_per_second = 0.063 +08/18/2024 17:37:18 - INFO - llamafactory.train.sft.trainer - Saving prediction results to ./results/inference/Llama2-7B/llama2_predict_1/generated_predictions.jsonl diff --git a/results/inference/Llama2-7B/llama2_predict_1_single.log b/results/inference/Llama2-7B/llama2_predict_1_single.log new file mode 100644 index 00000000..2853a6bf --- /dev/null +++ b/results/inference/Llama2-7B/llama2_predict_1_single.log @@ -0,0 +1,20 @@ +08/18/2024 17:42:59 - INFO - llamafactory.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: False, compute dtype: None +08/18/2024 17:43:00 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +eval example: +input_ids: +[1, 518, 25580, 29962, 29871, 30982, 31695, 31863, 31577, 30210, 30457, 30502, 31302, 30858, 30267, 518, 29914, 25580, 29962] +inputs: + [INST] 保持健康的三个提示。 [/INST] +08/18/2024 17:43:05 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:43:34 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:43:34 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +***** predict metrics ***** + predict_bleu-4 = 1.2952 + predict_model_preparation_time = 0.0029 + predict_rouge-1 = 12.8381 + predict_rouge-2 = 3.2551 + predict_rouge-l = 5.1021 + predict_runtime = 0:06:01.55 + predict_samples_per_second = 0.138 + predict_steps_per_second = 0.069 +08/18/2024 17:49:36 - INFO - llamafactory.train.sft.trainer - Saving prediction results to ./results/inference/Llama2-7B/llama2_predict_1_single/generated_predictions.jsonl diff --git a/results/inference/Llama2-7B/llama2_predict_2.log b/results/inference/Llama2-7B/llama2_predict_2.log new file mode 100644 index 00000000..f80b641d --- /dev/null +++ b/results/inference/Llama2-7B/llama2_predict_2.log @@ -0,0 +1,58 @@ +08/18/2024 17:37:25 - INFO - llamafactory.cli - Initializing distributed tasks at: 127.0.0.1:28639 +08/18/2024 17:37:31 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:31 - INFO - llamafactory.hparams.parser - Process rank: 1, device: cuda:1, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:32 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:32 - INFO - llamafactory.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:32 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:32 - INFO - llamafactory.hparams.parser - Process rank: 6, device: cuda:6, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:33 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:33 - INFO - llamafactory.hparams.parser - Process rank: 2, device: cuda:2, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:33 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:33 - INFO - llamafactory.hparams.parser - Process rank: 3, device: cuda:3, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:33 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:33 - INFO - llamafactory.hparams.parser - Process rank: 4, device: cuda:4, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:33 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:37:33 - INFO - llamafactory.hparams.parser - Process rank: 5, device: cuda:5, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:37:33 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:38:38 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:38:38 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:38:38 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:38:38 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:38:38 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:38:38 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +eval example: +input_ids: +[1, 518, 25580, 29962, 29871, 30982, 31695, 31863, 31577, 30210, 30457, 30502, 31302, 30858, 30267, 518, 29914, 25580, 29962] +inputs: + [INST] 保持健康的三个提示。 [/INST] +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:38:44 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:39:22 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:22 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:39:23 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:23 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:39:25 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:25 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:39:25 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:25 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:25 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:39:25 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:39:25 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:25 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:39:25 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:39:25 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +***** predict metrics ***** + predict_bleu-4 = 1.4552 + predict_model_preparation_time = 0.003 + predict_rouge-1 = 11.0984 + predict_rouge-2 = 3.0446 + predict_rouge-l = 5.0312 + predict_runtime = 0:01:03.40 + predict_samples_per_second = 0.789 + predict_steps_per_second = 0.063 +08/18/2024 17:40:28 - INFO - llamafactory.train.sft.trainer - Saving prediction results to ./results/inference/Llama2-7B/llama2_predict_2/generated_predictions.jsonl diff --git a/results/inference/Llama2-7B/llama2_predict_2_single.log b/results/inference/Llama2-7B/llama2_predict_2_single.log new file mode 100644 index 00000000..ffb6573b --- /dev/null +++ b/results/inference/Llama2-7B/llama2_predict_2_single.log @@ -0,0 +1,20 @@ +08/18/2024 17:49:41 - INFO - llamafactory.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: False, compute dtype: None +08/18/2024 17:49:42 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +eval example: +input_ids: +[1, 518, 25580, 29962, 29871, 30982, 31695, 31863, 31577, 30210, 30457, 30502, 31302, 30858, 30267, 518, 29914, 25580, 29962] +inputs: + [INST] 保持健康的三个提示。 [/INST] +08/18/2024 17:49:47 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:50:16 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:50:16 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +***** predict metrics ***** + predict_bleu-4 = 1.2952 + predict_model_preparation_time = 0.003 + predict_rouge-1 = 12.8381 + predict_rouge-2 = 3.2551 + predict_rouge-l = 5.1021 + predict_runtime = 0:06:01.42 + predict_samples_per_second = 0.138 + predict_steps_per_second = 0.069 +08/18/2024 17:56:18 - INFO - llamafactory.train.sft.trainer - Saving prediction results to ./results/inference/Llama2-7B/llama2_predict_2_single/generated_predictions.jsonl diff --git a/results/inference/Llama2-7B/llama2_predict_3.log b/results/inference/Llama2-7B/llama2_predict_3.log new file mode 100644 index 00000000..d7e7b8e7 --- /dev/null +++ b/results/inference/Llama2-7B/llama2_predict_3.log @@ -0,0 +1,58 @@ +08/18/2024 17:40:35 - INFO - llamafactory.cli - Initializing distributed tasks at: 127.0.0.1:24151 +08/18/2024 17:40:42 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:42 - INFO - llamafactory.hparams.parser - Process rank: 2, device: cuda:2, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:43 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:43 - INFO - llamafactory.hparams.parser - Process rank: 5, device: cuda:5, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:43 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:43 - INFO - llamafactory.hparams.parser - Process rank: 6, device: cuda:6, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:43 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:43 - INFO - llamafactory.hparams.parser - Process rank: 4, device: cuda:4, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:43 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:43 - INFO - llamafactory.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:43 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:43 - INFO - llamafactory.hparams.parser - Process rank: 1, device: cuda:1, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:43 - WARNING - llamafactory.hparams.parser - `ddp_find_unused_parameters` needs to be set as False for LoRA in DDP training. +08/18/2024 17:40:43 - INFO - llamafactory.hparams.parser - Process rank: 3, device: cuda:3, n_gpu: 1, distributed training: True, compute dtype: None +08/18/2024 17:40:44 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:40:54 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:40:54 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:40:54 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:40:54 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:40:54 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +08/18/2024 17:40:54 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +eval example: +input_ids: +[1, 518, 25580, 29962, 29871, 30982, 31695, 31863, 31577, 30210, 30457, 30502, 31302, 30858, 30267, 518, 29914, 25580, 29962] +inputs: + [INST] 保持健康的三个提示。 [/INST] +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:09 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:41:48 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:48 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:41:48 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:48 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:48 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:41:48 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:41:48 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:48 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:41:48 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:48 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:41:48 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:48 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +08/18/2024 17:41:49 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:41:49 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +***** predict metrics ***** + predict_bleu-4 = 1.4552 + predict_model_preparation_time = 0.0031 + predict_rouge-1 = 11.0984 + predict_rouge-2 = 3.0446 + predict_rouge-l = 5.0312 + predict_runtime = 0:01:03.96 + predict_samples_per_second = 0.782 + predict_steps_per_second = 0.063 +08/18/2024 17:42:52 - INFO - llamafactory.train.sft.trainer - Saving prediction results to ./results/inference/Llama2-7B/llama2_predict_3/generated_predictions.jsonl diff --git a/results/inference/Llama2-7B/llama2_predict_3_single.log b/results/inference/Llama2-7B/llama2_predict_3_single.log new file mode 100644 index 00000000..b425c65b --- /dev/null +++ b/results/inference/Llama2-7B/llama2_predict_3_single.log @@ -0,0 +1,20 @@ +08/18/2024 17:56:23 - INFO - llamafactory.hparams.parser - Process rank: 0, device: cuda:0, n_gpu: 1, distributed training: False, compute dtype: None +08/18/2024 17:56:24 - INFO - llamafactory.data.loader - Loading dataset llamafactory/alpaca_gpt4_zh... +eval example: +input_ids: +[1, 518, 25580, 29962, 29871, 30982, 31695, 31863, 31577, 30210, 30457, 30502, 31302, 30858, 30267, 518, 29914, 25580, 29962] +inputs: + [INST] 保持健康的三个提示。 [/INST] +08/18/2024 17:56:29 - INFO - llamafactory.model.patcher - Using KV cache for faster generation. +08/18/2024 17:56:58 - INFO - llamafactory.model.model_utils.attention - Using torch SDPA for faster training and inference. +08/18/2024 17:56:58 - INFO - llamafactory.model.loader - all params: 6,738,415,616 +***** predict metrics ***** + predict_bleu-4 = 1.2952 + predict_model_preparation_time = 0.003 + predict_rouge-1 = 12.8381 + predict_rouge-2 = 3.2551 + predict_rouge-l = 5.1021 + predict_runtime = 0:06:01.71 + predict_samples_per_second = 0.138 + predict_steps_per_second = 0.069 +08/18/2024 18:03:00 - INFO - llamafactory.train.sft.trainer - Saving prediction results to ./results/inference/Llama2-7B/llama2_predict_3_single/generated_predictions.jsonl