返回模型

说明文档

任务: token-classification
后端: sagemaker-training
后端参数: {'instance_type': 'ml.m5.2xlarge', 'supported_instructions': 'avx512'}
评估样本数量: 全部数据集

固定参数:

数据集: [{'path': 'conll2003', 'eval_split': 'validation', 'data_keys': {'primary': 'tokens'}, 'ref_keys': ['ner_tags'], 'name': None, 'calibration_split': 'train'}]
name_or_path: elastic/distilbert-base-uncased-finetuned-conll03-english
from_transformers: True
待量化算子: ['Add', 'MatMul']
校准:
- 方法: percentile
- 校准样本数: 128
- 校准直方图百分位: 99.999

基准测试参数:

框架: onnxruntime, pytorch
量化方法: dynamic, static
节点排除: [], ['layernorm', 'gelu', 'residual', 'gather', 'softmax']
逐通道量化: False, True
框架参数: {'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}, {}
reduce_range: True, False
应用量化: True, False

评估

非时间指标

框架	量化方法	节点排除	逐通道量化	框架参数	reduce_range	应用量化		整体精确率		整体召回率		整体F1值		整体准确率
`onnxruntime`	`None`	`None`	`None`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`None`	`False`	\|	0.936	\|	0.944	\|	0.940	\|	0.988
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.935	\|	0.943	\|	0.939	\|	0.988
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.926	\|	0.931	\|	0.929	\|	0.987
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.000	\|	0.000	\|	0.000	\|	0.833
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.934	\|	0.944	\|	0.939	\|	0.988
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.935	\|	0.943	\|	0.939	\|	0.988
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.926	\|	0.931	\|	0.929	\|	0.987
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.000	\|	0.000	\|	0.000	\|	0.833
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.934	\|	0.944	\|	0.939	\|	0.988
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.913	\|	0.792	\|	0.848	\|	0.969
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.913	\|	0.792	\|	0.848	\|	0.969
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.000	\|	0.000	\|	0.000	\|	0.833
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.896	\|	0.783	\|	0.836	\|	0.968
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.925	\|	0.844	\|	0.883	\|	0.975
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.925	\|	0.844	\|	0.883	\|	0.975
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	0.045	\|	0.004	\|	0.008	\|	0.825
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	0.922	\|	0.839	\|	0.879	\|	0.975
`pytorch`	`None`	`None`	`None`	`{}`	`None`	`None`	\|	0.936	\|	0.944	\|	0.940	\|	0.988

时间指标

时间基准测试每个配置运行15秒。

以下是批量大小 = 1，输入长度 = 32 的时间指标。

框架	量化方法	节点排除	逐通道量化	框架参数	reduce_range	应用量化		平均延迟 (ms)		吞吐量 (/s)
`onnxruntime`	`None`	`None`	`None`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`None`	`False`	\|	14.22	\|	70.33
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	10.22	\|	97.87
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	10.16	\|	98.47
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	10.52	\|	95.07
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	10.70	\|	93.47
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	10.22	\|	97.87
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	10.24	\|	97.67
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	10.36	\|	96.53
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	10.50	\|	95.27
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	10.98	\|	91.07
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	11.31	\|	88.47
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	11.23	\|	89.07
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	11.48	\|	87.20
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	13.54	\|	73.87
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	13.74	\|	72.80
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	13.80	\|	72.53
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	14.08	\|	71.07
`pytorch`	`None`	`None`	`None`	`{}`	`None`	`None`	\|	31.23	\|	32.07

以下是批量大小 = 1，输入长度 = 64 的时间指标。

框架	量化方法	节点排除	逐通道量化	框架参数	reduce_range	应用量化		平均延迟 (ms)		吞吐量 (/s)
`onnxruntime`	`None`	`None`	`None`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`None`	`False`	\|	24.52	\|	40.80
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	18.47	\|	54.20
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	18.53	\|	54.00
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	18.85	\|	53.07
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	19.14	\|	52.27
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	18.50	\|	54.07
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	18.50	\|	54.07
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	18.69	\|	53.53
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	19.46	\|	51.40
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	20.42	\|	49.00
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	19.91	\|	50.27
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	20.20	\|	49.53
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	20.74	\|	48.27
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	24.91	\|	40.20
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	24.35	\|	41.13
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	24.99	\|	40.07
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	24.95	\|	40.13
`pytorch`	`None`	`None`	`None`	`{}`	`None`	`None`	\|	41.31	\|	24.27

以下是批量大小 = 1，输入长度 = 128 的时间指标。

框架	量化方法	节点排除	逐通道量化	框架参数	reduce_range	应用量化		平均延迟 (ms)		吞吐量 (/s)
`onnxruntime`	`None`	`None`	`None`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`None`	`False`	\|	46.79	\|	21.40
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	35.84	\|	27.93
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	35.07	\|	28.53
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	35.71	\|	28.00
`onnxruntime`	`dynamic`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	35.91	\|	27.87
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	35.42	\|	28.27
`onnxruntime`	`dynamic`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	35.22	\|	28.40
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	35.51	\|	28.20
`onnxruntime`	`dynamic`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	35.90	\|	27.87
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	39.88	\|	25.13
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	39.27	\|	25.47
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	39.37	\|	25.40
`onnxruntime`	`static`	`['layernorm', 'gelu', 'residual', 'gather', 'softmax']`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	39.16	\|	25.60
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	44.43	\|	22.53
`onnxruntime`	`static`	`[]`	`False`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	46.13	\|	21.73
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`False`	`True`	\|	45.48	\|	22.00
`onnxruntime`	`static`	`[]`	`True`	`{'opset': 13, 'optimization_level': 1, 'intra_op_num_threads': 4}`	`True`	`True`	\|	45.82	\|	21.87
`pytorch`	`None`	`None`	`None`	`{}`	`None`	`None`	\|	53.93	\|	18.60

fxmarty/20220911-h13m58s51_conll2003_distilbert_quantization

作者 fxmarty

token-classification

↓ 0 ♥ 0

创建时间: 2022-09-11 15:14:40+00:00

更新时间: 2022-09-11 15:15:44+00:00

在 Hugging Face 上查看

文件 (105)

.gitattributes

20220911-h14m00s39_0/model.onnx ONNX

20220911-h14m00s39_0/ort_config.json

20220911-h14m00s39_0/quantized_model.onnx ONNX

20220911-h14m00s39_0/results.json

20220911-h14m02s15_1/model.onnx ONNX

20220911-h14m02s15_1/ort_config.json

20220911-h14m02s15_1/quantized_model.onnx ONNX

20220911-h14m02s15_1/results.json

20220911-h14m09s38_2/augmented_model.onnx ONNX

20220911-h14m09s38_2/calibration_histograms.npy

20220911-h14m09s38_2/model.onnx ONNX

20220911-h14m09s38_2/ort_config.json

20220911-h14m09s38_2/quantized_model.onnx ONNX

20220911-h14m09s38_2/results.json

20220911-h14m16s58_3/augmented_model.onnx ONNX

20220911-h14m16s58_3/calibration_histograms.npy

20220911-h14m16s58_3/model.onnx ONNX

20220911-h14m16s58_3/ort_config.json

20220911-h14m16s58_3/quantized_model.onnx ONNX

20220911-h14m16s58_3/results.json

20220911-h14m24s25_4/augmented_model.onnx ONNX

20220911-h14m24s25_4/calibration_histograms.npy

20220911-h14m24s25_4/model.onnx ONNX

20220911-h14m24s25_4/ort_config.json

20220911-h14m24s25_4/quantized_model.onnx ONNX

20220911-h14m24s25_4/results.json

20220911-h14m25s53_5/model.onnx ONNX

20220911-h14m25s53_5/ort_config.json

20220911-h14m25s53_5/quantized_model.onnx ONNX

20220911-h14m25s53_5/results.json

20220911-h14m33s27_6/augmented_model.onnx ONNX

20220911-h14m33s27_6/calibration_histograms.npy

20220911-h14m33s27_6/model.onnx ONNX

20220911-h14m33s27_6/ort_config.json

20220911-h14m33s27_6/quantized_model.onnx ONNX

20220911-h14m33s27_6/results.json

20220911-h14m40s51_7/augmented_model.onnx ONNX

20220911-h14m40s51_7/calibration_histograms.npy

20220911-h14m40s51_7/model.onnx ONNX

20220911-h14m40s51_7/ort_config.json

20220911-h14m40s51_7/quantized_model.onnx ONNX

20220911-h14m40s51_7/results.json

20220911-h14m42s27_8/model.onnx ONNX

20220911-h14m42s27_8/ort_config.json

20220911-h14m42s27_8/quantized_model.onnx ONNX

20220911-h14m42s27_8/results.json

20220911-h14m44s03_9/model.onnx ONNX

20220911-h14m44s03_9/ort_config.json

20220911-h14m44s03_9/quantized_model.onnx ONNX

20220911-h14m44s03_9/results.json

20220911-h14m45s31_10/model.onnx ONNX

20220911-h14m45s31_10/ort_config.json

20220911-h14m45s31_10/quantized_model.onnx ONNX

20220911-h14m45s31_10/results.json

20220911-h14m46s59_11/model.onnx ONNX

20220911-h14m46s59_11/ort_config.json

20220911-h14m46s59_11/quantized_model.onnx ONNX

20220911-h14m46s59_11/results.json

20220911-h14m48s28_12/model.onnx ONNX

20220911-h14m48s28_12/ort_config.json

20220911-h14m48s28_12/quantized_model.onnx ONNX

20220911-h14m48s28_12/results.json

20220911-h14m55s55_13/augmented_model.onnx ONNX

20220911-h14m55s55_13/calibration_histograms.npy

20220911-h14m55s55_13/model.onnx ONNX

20220911-h14m55s55_13/ort_config.json

20220911-h14m55s55_13/quantized_model.onnx ONNX

20220911-h14m55s55_13/results.json

20220911-h15m03s26_14/augmented_model.onnx ONNX

20220911-h15m03s26_14/calibration_histograms.npy

20220911-h15m03s26_14/model.onnx ONNX

20220911-h15m03s26_14/ort_config.json

20220911-h15m03s26_14/quantized_model.onnx ONNX

20220911-h15m03s26_14/results.json

20220911-h15m10s51_15/augmented_model.onnx ONNX

20220911-h15m10s51_15/calibration_histograms.npy

20220911-h15m10s51_15/model.onnx ONNX

20220911-h15m10s51_15/ort_config.json

20220911-h15m10s51_15/quantized_model.onnx ONNX

20220911-h15m10s51_15/results.json

20220911-h15m12s25_16/model.onnx ONNX

20220911-h15m12s25_16/results.json

20220911-h15m14s39_17/results.json

README.md

runs.json

tensorboard/1662909285.3827937/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.1

tensorboard/1662909285.3842196/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.2

tensorboard/1662909285.3854108/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.3

tensorboard/1662909285.3865986/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.4

tensorboard/1662909285.3881178/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.5

tensorboard/1662909285.389244/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.6

tensorboard/1662909285.390423/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.7

tensorboard/1662909285.391528/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.8

tensorboard/1662909285.3926113/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.9

tensorboard/1662909285.3937325/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.10

tensorboard/1662909285.3948941/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.11

tensorboard/1662909285.3960202/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.12

tensorboard/1662909285.3971512/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.13

tensorboard/1662909285.398288/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.14

tensorboard/1662909285.399385/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.15

tensorboard/1662909285.402039/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.16

tensorboard/1662909285.4033327/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.17

tensorboard/1662909285.4047217/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.18

tensorboard/events.out.tfevents.1662909285.ip-10-0-209-21.ec2.internal.1.0