Add hls4ml example

Author: axiotisk

vitis_accelerator_test/tuto_vitis_stream.py

@@ -0,0 +1,51 @@
+import hls4ml
+from hls4ml.backends import VitisAcceleratorConfig
+
+# Fetch a keras model from our example repository
+# This will download our example model to your working directory and return an example configuration file
+config = hls4ml.utils.fetch_example_model('KERAS_3layer.json')
+
+# You can print it to see some default parameters
+#print(config)
+
+# Default config:
+#{
+#    'OutputDir': 'my-hls-test',
+#    'ProjectName': 'myproject',
+#    'Backend': 'Vivado',
+#    'Part': 'xcku115-flvb2104-2-i',
+#    'ClockPeriod': 5,
+#    'IOType': 'io_parallel',
+#    'HLSConfig': {
+#        'Model': {
+#            'Precision': 'ap_fixed<16,6>',
+#            'ReuseFactor': 1
+#        }
+#    },
+#    'KerasJson': 'KERAS_3layer.json',
+#    'KerasH5': 'KERAS_3layer_weights.h5'
+#}
+
+config['OutputDir'] = 'prj_tuto_vitis_stream'
+config['ProjectName'] = 'PrjTutoVitisStream'
+config['Backend'] = 'VitisAccelerator'
+config['Part'] = 'xcu50-fsvh2104-2-e'
+config['IOType'] = 'io_stream'
+config['AcceleratorConfig'] = {}  # Initialize AcceleratorConfig as a dictionary
+config['AcceleratorConfig']['Platform'] = 'xilinx_u50_gen3x16_xdma_5_202210_1'
+
+print("[K] Print config after setting up backend:")
+print(config)
+
+# Convert it to a hls project
+hls_model = hls4ml.converters.keras_to_hls(config)
+
+## Print full list of example model if you want to explore more
+#hls4ml.utils.fetch_example_list()
+
+# Use Vivado HLS to synthesize the model
+# This might take several minutes
+hls_model.build()
+
+# Print out the report if you want
+hls4ml.report.read_vivado_report(' prj_tuto_vitis_stream')