Create julia examples.

examples/julia/tasks/audio_classification/audio_classification.jl

@@ -0,0 +1,35 @@
+using Bytez
+using Base64
+using HTTP
+
+
+# Function to get the audio as base64-encoded string
+function get_base64_audio(url::String)::String
+	response = HTTP.get(url)
+	return base64encode(response.body)
+end
+
+# Get the base64 encoded audio from the URL
+input_audio_base64 = get_base64_audio(
+	"https://huggingface.co/datasets/huggingfacejs/tasks/resolve/main/audio-classification/audio.wav",
+)
+
+# Initialize the Bytez client with your API key
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+# Load the model
+model = client.model("aaraki/wav2vec2-base-finetuned-ks")
+
+model.load()
+
+# Run inference on the model
+result = model.run(Dict("b64AudioBufferWav" => input_audio_base64))
+
+# Process and print the results
+label_objects = result["output"]
+
+for label_object in label_objects
+	score = label_object["score"]
+	label = label_object["label"]
+	println("Score: $score, Label: $label")
+end

examples/julia/tasks/automatic_speech_recognition/automatic_speech_recognition.jl

@@ -0,0 +1,32 @@
+using Bytez
+using Base64
+using HTTP
+
+function get_base64_audio(url::String)::String
+	response = HTTP.get(url)
+	return base64encode(response.body)
+end
+
+input_audio_base64 = get_base64_audio(
+	"https://huggingface.co/datasets/huggingfacejs/tasks/resolve/main/audio-classification/audio.wav",
+)
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("facebook/data2vec-audio-base-960h")
+
+model.load()
+
+result = model.run(Dict("b64AudioBufferWav" => input_audio_base64))
+
+output = result["output"]
+
+# depending on the model, there may be additional props returned
+println(output)
+
+text = output["text"]
+
+println("Inference is: $text")
+
+a = 2
+

examples/julia/tasks/chat_models/chat_models.jl

@@ -0,0 +1,28 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("microsoft/Phi-3-mini-4k-instruct")
+
+model.load()
+
+messages = [
+	Dict("role" => "system", "content" => "You are a friendly chatbot"),
+	Dict("role" => "user", "content" => "What is the capital of England?"),
+]
+
+result = model.run(messages, Dict("max_length" => 100))
+
+output = result["output"]
+
+generated_text = output[1]["generated_text"]
+
+for message in generated_text
+	# depending on the model, there may be additional props returned
+	println(message)
+
+	content = message["content"]
+	role = message["role"]
+
+	println(Dict("content" => content, "role" => role))
+end

examples/julia/tasks/depth_estimation/depth_estimation.jl

@@ -0,0 +1,50 @@
+using Bytez
+using Base64
+using HTTP
+using Printf
+
+WORKING_DIR = dirname(@__FILE__)
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("vinvino02/glpn-nyu")
+
+model.load()
+
+input_image_url = "https://as1.ftcdn.net/v2/jpg/03/03/55/82/1000_F_303558268_YNUQp9NNMTE0X4zrj314mbWcDHd1pZPD.jpg"
+
+# Run the model with the input image
+result = model.run(input_image_url)
+
+output = result["output"]
+
+#  depending on the model, there may be additional props returned
+depth_png = output["depth_png"]
+formatted_predicted_depth_array = output["formatted_predicted_depth_array"]
+
+##### Decode and save the image #####
+img_buffer = base64decode(depth_png)
+
+image_path = joinpath(WORKING_DIR, "testImage.png")
+open(image_path, "w") do f
+	write(f, img_buffer)
+end
+
+# Write the original image for comparison
+original_image_path = joinpath(WORKING_DIR, "originalImage.jpg")
+response = HTTP.get(input_image_url)
+
+open(original_image_path, "w") do f
+	write(f, response.body)
+end
+
+println("Wrote the original image to: ", original_image_path)
+println("Wrote the inference image to: ", image_path)
+
+##### 2D depth map, object representation of the pixel values for the depth map #####
+rows = formatted_predicted_depth_array
+for (j, row) in enumerate(rows)
+	for (i, pixel) in enumerate(row)
+		# insert code here if you need these values directly
+	end
+end

examples/julia/tasks/document_question_answering/document_question_answering.jl

@@ -0,0 +1,32 @@
+using Bytez
+using Printf
+
+WORKING_DIR = dirname(@__FILE__)
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("cloudqi/CQI_Visual_Question_Awnser_PT_v0")
+
+model.load()
+
+input_data = Dict(
+	"image" => "https://templates.invoicehome.com/invoice-template-us-neat-750px.png",
+	"question" => "What's the total cost?",
+)
+
+result = model.run(input_data)
+
+output = result["output"]
+
+#  depending on the model, there may be additional props returned
+println(output)
+
+output_object = output[1]
+
+answer = output_object["answer"]
+score = output_object["score"]
+start = output_object["start"]
+# end is a reserved keyword in julia
+_end = output_object["end"]
+
+println(Dict("answer" => answer, "score" => score, "start" => start, "end" => _end))

examples/julia/tasks/feature_extraction/feature_extraction.jl

@@ -0,0 +1,20 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("Salesforce/SFR-Embedding-2_R")
+
+model.load()
+
+input_text = "Your text for feature extraction goes here..."
+
+result = model.run(input_text)
+
+output = result["output"]
+
+#  depending on the model, there may be additional props returned
+println(output)
+
+embedding = output[1]
+
+println(embedding)

examples/julia/tasks/fill_mask/fill_mask.jl

@@ -0,0 +1,25 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("almanach/camembert-base")
+
+model.load()
+
+input_text = "The capital of France is <mask>."
+
+result = model.run(input_text)
+
+sequence_objects = result["output"]
+
+for sequence_object in sequence_objects
+	#  depending on the model, there may be additional props returned
+	println(sequence_object)
+
+	sequence = sequence_object["sequence"]
+	score = sequence_object["score"]
+	token = sequence_object["token"]
+	token_str = sequence_object["token_str"]
+
+	println(Dict("sequence" => sequence, "score" => score, "token" => token, "token_str" => token_str))
+end

examples/julia/tasks/function_calling_models/function_calling_models.jl

@@ -0,0 +1,56 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("Nexusflow/NexusRaven-V2-13B")
+
+model.load()
+
+input_text = "What's the weather like in Seattle right now?"
+
+options = Dict(
+	"params" => Dict(
+		"max_new_tokens" => 20,
+		"min_new_tokens" => 50,
+		"temperature" => 0.001,
+		"do_sample" => false),
+	"stream" => true,
+)
+
+prompt_template = """
+Function:
+def get_weather_data(coordinates):
+	\"\"\"
+	Fetches weather data from the Open-Meteo API for the given latitude and longitude.
+
+	Args:
+	coordinates (tuple): The latitude and longitude of the location.
+
+	Returns:
+	float: The current temperature in the coordinates you've asked for
+	\"""
+
+Function:
+def get_coordinates_from_city(city_name):
+	\"""
+	Fetches the latitude and longitude of a given city name using the Maps.co Geocoding API.
+
+	Args:
+	city_name (str): The name of the city.
+
+	Returns:
+	tuple: The latitude and longitude of the city.
+	\"""
+
+User Query: {query}<human_end>
+"""
+
+# Prepare the prompt with the user query
+prompt = replace(prompt_template, "{query}" => input_text)
+
+stream = model.run(prompt, options)
+
+while isopen(stream)
+	item = take!(stream)  # Take each item as it enters the channel
+	println(item)  # Print the item
+end

examples/julia/tasks/image_classification/image_classification.jl

@@ -0,0 +1,23 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("google/vit-base-patch16-224")
+
+model.load()
+
+input_image_url = "https://www.padoniavets.com/sites/default/files/field/image/cats-and-dogs.jpg"
+
+result = model.run(input_image_url)
+
+labelObjects = result["output"]
+
+for labelObject in labelObjects
+	#  depending on the model, there may be additional props returned
+	println(labelObject)
+
+	label = labelObject["label"]
+	score = labelObject["score"]
+
+	println(Dict("label" => label, "score" => score))
+end

examples/julia/tasks/image_feature_extraction/image_feature_extraction.jl

@@ -0,0 +1,20 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("nomic-ai/nomic-embed-vision-v1")
+
+model.load()
+
+input_image_url = "https://as1.ftcdn.net/v2/jpg/03/03/55/82/1000_F_303558268_YNUQp9NNMTE0X4zrj314mbWcDHd1pZPD.jpg"
+
+result = model.run(input_image_url)
+
+output = result["output"]
+
+#  depending on the model, there may be additional props returned
+println(output)
+
+embedding = output[1]
+
+println(embedding)

examples/julia/tasks/image_segmentation/image_segmentation.jl

@@ -0,0 +1,34 @@
+using Bytez
+using Base64
+using Printf
+
+WORKING_DIR = dirname(@__FILE__)
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+model = client.model("sayeed99/segformer-b3-fashion")
+
+model.load()
+
+input_image_url = "https://ocean.si.edu/sites/default/files/styles/3_2_largest/public/2023-11/Screen_Shot_2018-04-16_at_1_42_56_PM.png.webp?itok=Icvi-ek9"
+
+result = model.run(input_image_url)
+
+mask_objects = result["output"]
+
+for (index, mask_object) in enumerate(mask_objects)
+	#  depending on the model, there may be additional props returned
+	println(mask_object)
+
+	label = mask_object["label"]
+	score = mask_object["score"]
+	mask_png = mask_object["mask_png"]
+
+	println(Dict("label" => label, "score" => score))
+
+	mask_png_buffer = base64decode(mask_png)
+
+	open("$(WORKING_DIR)/mask-$index.png", "w") do file
+		write(file, mask_png_buffer)
+	end
+end

examples/julia/tasks/image_to_text/image_to_text.jl

@@ -0,0 +1,20 @@
+using Bytez
+
+client = Bytez.init("YOUR BYTEZ KEY HERE")
+
+input_image_url = "https://as1.ftcdn.net/v2/jpg/03/03/55/82/1000_F_303558268_YNUQp9NNMTE0X4zrj314mbWcDHd1pZPD.jpg"
+
+model = client.model("Salesforce/blip-image-captioning-base")
+
+model.load()
+
+result = model.run(input_image_url)
+
+#  depending on the model, there may be additional props returned
+output = result["output"]
+
+println(output)
+
+generated_text = output[1]["generated_text"]
+
+println(generated_text)