Update example model export directories

Author: deliahu

examples/sentiment-analysis/README.md

@@ -10,7 +10,7 @@ This example shows how to deploy a sentiment analysis classifier trained using [
 
 - kind: api
   name: classifier
-  model: s3://cortex-examples/sentiment
+  model: s3://cortex-examples/sentiment-analysis/bert
   request_handler: sentiment.py
 ```
 

examples/sentiment-analysis/bert.ipynb

@@ -933,7 +933,7 @@
       "source": [
         "AWS_ACCESS_KEY_ID = \"\" #@param {type:\"string\"}\n",
         "AWS_SECRET_ACCESS_KEY = \"\" #@param {type:\"string\"}\n",
-        "S3_UPLOAD_PATH = \"s3://my-bucket/bert\" #@param {type:\"string\"}\n",
+        "S3_UPLOAD_PATH = \"s3://my-bucket/sentiment-analysis/bert\" #@param {type:\"string\"}\n",
         "\n",
         "import sys\n",
         "import re\n",

examples/sentiment-analysis/cortex.yaml

@@ -3,7 +3,7 @@
 
 - kind: api
   name: classifier
-  model: s3://cortex-examples/bert
+  model: s3://cortex-examples/sentiment-analysis/bert
   request_handler: handler.py
   tracker:
     model_type: classification

examples/text-generator/README.md

@@ -10,13 +10,13 @@ This example shows how to deploy OpenAI's GPT-2 model as a service on AWS.
 
 - kind: api
   name: generator
-  model: s3://cortex-examples/gpt-2/774
+  model: s3://cortex-examples/text-generator/gpt-2/124M
   request_handler: encoder.py
   compute:
     gpu: 1
 ```
 
-A `deployment` specifies a set of resources that are deployed as a single unit. An `api` makes a model available as a web service that can serve real-time predictions. This configuration will download the 774M GPT-2 model from the `cortex-examples` S3 bucket, preprocess the payload and postprocess the inference with functions defined in `encoder.py` and deploy each replica of the API on 1 GPU.
+A `deployment` specifies a set of resources that are deployed as a single unit. An `api` makes a model available as a web service that can serve real-time predictions. This configuration will download the 124M GPT-2 model from the `cortex-examples` S3 bucket, preprocess the payload and postprocess the inference with functions defined in `encoder.py` and deploy each replica of the API on 1 GPU.
 
 ## Add request handling
 

examples/text-generator/cortex.yaml

@@ -3,7 +3,7 @@
 
 - kind: api
   name: generator
-  model: s3://cortex-examples/gpt-2/124M
+  model: s3://cortex-examples/text-generator/gpt-2/124M
   request_handler: handler.py
   compute:
     gpu: 1

examples/text-generator/encoder.py

@@ -105,9 +105,11 @@ def decode(self, tokens):
 
 def get_encoder():
     s3 = boto3.client("s3")
-    encoder = json.load(s3.get_object(Bucket="cortex-examples", Key="gpt-2/encoder.json")["Body"])
+    encoder = json.load(
+        s3.get_object(Bucket="cortex-examples", Key="text-generator/gpt-2/encoder.json")["Body"]
+    )
     bpe_data = (
-        s3.get_object(Bucket="cortex-examples", Key="gpt-2/vocab.bpe")["Body"]
+        s3.get_object(Bucket="cortex-examples", Key="text-generator/gpt-2/vocab.bpe")["Body"]
         .read()
         .decode("utf-8")
     )

examples/text-generator/gpt-2.ipynb

@@ -279,7 +279,7 @@
       "source": [
         "AWS_ACCESS_KEY_ID = \"\" #@param {type:\"string\"}\n",
         "AWS_SECRET_ACCESS_KEY = \"\" #@param {type:\"string\"}\n",
-        "S3_UPLOAD_PATH = \"s3://my-bucket/gpt-2\" #@param {type:\"string\"}\n",
+        "S3_UPLOAD_PATH = \"s3://my-bucket/text-generator/gpt-2\" #@param {type:\"string\"}\n",
         "\n",
         "import sys\n",
         "import re\n",