GCDEC/Tensorflow/Notes - Revision history

Admin at 10:21, 7 January 2018

2018-01-07T10:21:03Z

2018-01-07T10:20:47Z

2018-01-07T10:20:34Z

← Older revision		Revision as of 10:20, 7 January 2018
Line 2:		Line 2:
	* [[GCDEC/Building Tensorflow/Notes]]		* [[GCDEC/Building Tensorflow/Notes]]
	* [[GCDEC/Deploying Tensorflow/Notes]]		* [[GCDEC/Deploying Tensorflow/Notes]]
			* [[GCDEC/Engineering Tensorflow/Notes]]



	[[Category:Google Cloud]]		[[Category:Google Cloud]]
	[[Category:Data Engineering]]		[[Category:Data Engineering]]

2017-10-18T09:15:35Z

← Older revision		Revision as of 09:15, 18 October 2017
Line 1:		Line 1:
	~~=References=~~		See the two sub-pages:
			* [[GCDEC/Building Tensorflow/Notes]]
			* [[GCDEC/Deploying Tensorflow/Notes]]



	~~=Flags=~~

	[[Category:Google Cloud]]		[[Category:Google Cloud]]
	[[Category:Data Engineering]]		[[Category:Data Engineering]]

2017-10-18T07:07:15Z

Pick Up Here

2017-10-18T07:05:02Z

Serverless Machine Learning with TensorFlow on Google Cloud Platform

2017-10-18T07:02:04Z

Cloud ML Engine for Prediction

← Older revision		Revision as of 07:02, 18 October 2017
Line 1,011:		Line 1,011:
	}]		}]

	# Now assemble the URL to which to send the model inputs		# Now assemble the URL to which to send the model inputs:
			# Set the following information:
			# - name of the project
			# - name of the model
			# - name of the version
	parent = 'projects/%s/models/%s/versions/%s' % ( 'cloud-training-demos', 'taxifare', 'v1' )		parent = 'projects/%s/models/%s/versions/%s' % ( 'cloud-training-demos', 'taxifare', 'v1' )

Line 1,018:		Line 1,022:
	name = parent		name = parent
	}).execute()		}).execute()
			</pre>

			Recall that we specified the model and version number when we ran <code>gcloud ml-engine models create</code> and <code>gcloud ml-engine versions create</code>:

			<pre>
			MODEL_NAME="taxifare"
			MODEL_VERSION="v1"
			MODEL_LOCATION="gs://${BUCKET}/taxifare/smallinput/taxi_trained/export/Servo/..."
	</pre>		</pre>

2017-10-18T07:00:19Z

Client interaction with model predictions

← Older revision		Revision as of 07:00, 18 October 2017
Line 1,011:		Line 1,011:
	}]		}]

	# Now ~~construct~~ the ~~name of~~ the model		# Now assemble the URL to which to send the model inputs
	parent = 'projects/%s/models/%s/versions/%s' % ( 'cloud-training-demos', 'taxifare', 'v1' )		parent = 'projects/%s/models/%s/versions/%s' % ( 'cloud-training-demos', 'taxifare', 'v1' )

			# Make the API request (call the predict function)
	response = api.projects().predict( body = { 'instances' : request_data,		response = api.projects().predict( body = { 'instances' : request_data,
	name = parent		name = parent

2017-10-18T06:59:40Z

Cloud ML Engine for Prediction

@@ Line 951: / Line 951: @@
      eval_metrics = model.get_eval_metrics(),
+)
 </pre>

2017-10-18T06:45:52Z

Cloud ML Engine for Prediction

← Older revision		Revision as of 06:45, 18 October 2017
Line 942:		Line 942:
	train_input_fn = model.generate_csv_input_fn( train_data_paths, ... ),		train_input_fn = model.generate_csv_input_fn( train_data_paths, ... ),
	eval_input_fn = model.generate_csv_input_fn( eval_data_paths, ... ),		eval_input_fn = model.generate_csv_input_fn( eval_data_paths, ... ),

			export_strategies = [saved_model_export_utils.make_export_strategy(
			model.serving_input_fn,
			default_output_alternative_key = None,
			exports_to_keep = 1
			)],

	eval_metrics = model.get_eval_metrics(),		eval_metrics = model.get_eval_metrics(),
	)		)