Updating model versions (#2244)

* Updating model versions

* Rename ft pipeline names in notebooks

* updating the model version and the registry

* changing the job name

---------

Co-authored-by: Pavan Manoj Jonnalagadda <pavanmanojj@microsoft.com>

cli/foundation-models/system/finetune/question-answering/extractive-qa-pipeline.yml

@@ -11,7 +11,7 @@ inputs:
 
   # specify the foundation model available in the azureml system registry
   mlflow_model_path: 
-    path: azureml://registries/azureml-preview/models/bert-based-uncased/versions/3
+    path: azureml://registries/azureml/models/bert-based-uncased/versions/3
     # huggingface_id: 'bert-base-uncased' # if you want to use a huggingface model, uncomment this line and comment the above line
 
   # map the dataset files to parameters
@@ -60,10 +60,9 @@ settings:
   force_rerun: true
 
 jobs:
-  extractive_qna_finetune_job:
+  question_answering_pipeline:
     type: pipeline
-    # component: azureml://registries/azureml-preview/components/question_answering_pipeline/versions/0.0.3
-    component: azureml://registries/azureml-preview/components/question_answering_pipeline/labels/latest
+    component: azureml://registries/azureml/components/question_answering_pipeline/labels/latest
     inputs:
       mlflow_model_path: ${{parent.inputs.mlflow_model_path}} 
 

cli/foundation-models/system/finetune/question-answering/extractive-qa.sh

@@ -18,7 +18,7 @@ gpus_per_node=2
 # This is the foundation model for finetuning
 model_name="bert-base-uncased"
 # using the latest version of the model - not working yet
-model_version=1
+model_version=3
 
 version=$(date +%s)
 finetuned_model_name=$model_name"-extractive-qna"
@@ -118,7 +118,7 @@ fi
 # need to switch to using latest version for model, currently blocked with a bug.
 # submit finetuning job
 parent_job_name=$( az ml job create --file ./extractive-qa-pipeline.yml $workspace_info --query name -o tsv --set \
-  jobs.extractive_qna_finetune_job.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
+  jobs.question_answering_pipeline.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
   inputs.compute_model_import=$compute_cluster \
   inputs.compute_preprocess=$compute_cluster \
   inputs.compute_finetune=$compute_cluster \

cli/foundation-models/system/finetune/summarization/news-summary-pipeline.yml

@@ -11,7 +11,7 @@ inputs:
 
   # specify the foundation model available in the azureml system registry
   mlflow_model_path: 
-    path: azureml://registries/azureml-preview/models/t5-small/versions/4
+    path: azureml://registries/azureml/models/t5-small/versions/3
     # huggingface_id: 't5-small' # if you want to use a huggingface model, uncomment this line and comment the above line
 
   # map the dataset files to parameters
@@ -50,10 +50,9 @@ settings:
   force_rerun: true
 
 jobs:
-  news_summary_finetune_job:
+  summarization_pipeline:
     type: pipeline
-    # component: azureml://registries/azureml-preview/components/summarization_pipeline/versions/0.0.3
-    component: azureml://registries/azureml-preview/components/summarization_pipeline/labels/latest
+    component: azureml://registries/azureml/components/summarization_pipeline/labels/latest
     inputs:
       mlflow_model_path: ${{parent.inputs.mlflow_model_path}} 
 

cli/foundation-models/system/finetune/summarization/news-summary.sh

@@ -18,7 +18,7 @@ gpus_per_node=2
 # This is the foundation model for finetuning
 model_name="t5-small"
 # using the latest version of the model - not working yet
-model_version=1
+model_version=3
 
 version=$(date +%s)
 finetuned_model_name=$model_name"-news-summary"
@@ -107,7 +107,7 @@ fi
 # need to switch to using latest version for model, currently blocked with a bug.
 # submit finetuning job
 parent_job_name=$( az ml job create --file ./news-summary-pipeline.yml $workspace_info --query name -o tsv --set \
-  jobs.news_summary_finetune_job.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
+  jobs.summarization_pipeline.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
   inputs.compute_model_import=$compute_cluster \
   inputs.compute_preprocess=$compute_cluster \
   inputs.compute_finetune=$compute_cluster \

cli/foundation-models/system/finetune/text-classification/emotion-detection-pipeline.yml

@@ -11,7 +11,7 @@ inputs:
 
   # specify the foundation model available in the azureml system registry
   mlflow_model_path: 
-    path: azureml://registries/azureml-preview/models/bert-based-uncased/versions/3
+    path: azureml://registries/azureml/models/bert-based-uncased/versions/3
     # huggingface_id: 'bert-base-uncased' # if you want to use a huggingface model, uncomment this line and comment the above line
 
   # map the dataset files to parameters
@@ -48,10 +48,9 @@ settings:
   force_rerun: true
 
 jobs:
-  emotion_detection_finetune_job:
+  text_classification_pipeline:
     type: pipeline
-    # component: azureml://registries/azureml-preview/components/text_classification_pipeline/versions/0.0.3
-    component: azureml://registries/azureml-preview/components/text_classification_pipeline/labels/latest
+    component: azureml://registries/azureml/components/text_classification_pipeline/labels/latest
     inputs:
       mlflow_model_path: ${{parent.inputs.mlflow_model_path}} 
 

cli/foundation-models/system/finetune/text-classification/emotion-detection.sh

@@ -18,7 +18,7 @@ gpus_per_node=2
 # This is the foundation model for finetuning
 model_name="bert-base-uncased"
 # using the latest version of the model - not working yet
-model_version=1
+model_version=3
 
 version=$(date +%s)
 finetuned_model_name=$model_name"-emotion-detection"
@@ -105,7 +105,7 @@ fi
 # need to switch to using latest version for model, currently blocked with a bug.
 # submit finetuning job
 parent_job_name=$( az ml job create --file ./emotion-detection-pipeline.yml $workspace_info --query name -o tsv --set \
-  jobs.emotion_detection_finetune_job.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
+  jobs.text_classification_pipeline.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
   inputs.compute_model_import=$compute_cluster \
   inputs.compute_preprocess=$compute_cluster \
   inputs.compute_finetune=$compute_cluster \

cli/foundation-models/system/finetune/token-classification/token-classification-pipeline.yml

@@ -11,7 +11,7 @@ inputs:
 
   # specify the foundation model available in the azureml system registry
   mlflow_model_path: 
-    path: azureml://registries/azureml-preview/models/bert-based-uncased/versions/3
+    path: azureml://registries/azureml/models/bert-based-uncased/versions/3
     # huggingface_id: 'bert-base-uncased' # if you want to use a huggingface model, uncomment this line and comment the above line
 
   # map the dataset files to parameters
@@ -48,10 +48,9 @@ settings:
   force_rerun: true
 
 jobs:
-  ner_finetune_job:
+  token_classification_pipeline:
     type: pipeline
-    # component: azureml://registries/azureml-preview/components/token_classification_pipeline/versions/0.0.3
-    component: azureml://registries/azureml-preview/components/token_classification_pipeline/labels/latest
+    component: azureml://registries/azureml/components/token_classification_pipeline/labels/latest
     inputs:
       mlflow_model_path: ${{parent.inputs.mlflow_model_path}} 
 

cli/foundation-models/system/finetune/token-classification/token-classification.sh

@@ -18,7 +18,7 @@ gpus_per_node=2
 # This is the foundation model for finetuning
 model_name="bert-base-uncased"
 # using the latest version of the model - not working yet
-model_version=1
+model_version=3
 
 version=$(date +%s)
 finetuned_model_name=$model_name"-ner"
@@ -105,7 +105,7 @@ fi
 # need to switch to using latest version for model, currently blocked with a bug.
 # submit finetuning job
 parent_job_name=$( az ml job create --file ./token-classification-pipeline.yml $workspace_info --query name -o tsv --set \
-  jobs.ner_finetune_job.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
+  jobs.token_classification_pipeline.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
   inputs.compute_model_import=$compute_cluster \
   inputs.compute_preprocess=$compute_cluster \
   inputs.compute_finetune=$compute_cluster \

cli/foundation-models/system/finetune/translation/translation-pipeline.yml

@@ -11,7 +11,7 @@ inputs:
 
   # specify the foundation model available in the azureml system registry
   mlflow_model_path: 
-    path: azureml://registries/azureml-preview/models/t5-small/versions/4
+    path: azureml://registries/azureml/models/t5-small/versions/3
     # huggingface_id: 't5-small' # if you want to use a huggingface model, uncomment this line and comment the above line
 
   # map the dataset files to parameters
@@ -50,10 +50,9 @@ settings:
   force_rerun: true
 
 jobs:
-  translation_en_ro_finetune_job:
+  translation_pipeline:
     type: pipeline
-    # component: azureml://registries/azureml-preview/components/translation_pipeline/versions/0.0.3
-    component: azureml://registries/azureml-preview/components/translation_pipeline/labels/latest
+    component: azureml://registries/azureml/components/translation_pipeline/labels/latest
     inputs:
       mlflow_model_path: ${{parent.inputs.mlflow_model_path}} 
 

cli/foundation-models/system/finetune/translation/translation.sh

@@ -19,7 +19,7 @@ gpus_per_node=2
 # This is the foundation model for finetuning
 model_name="t5-small"
 # using the latest version of the model - not working yet
-model_version=1
+model_version=3
 
 version=$(date +%s)
 finetuned_model_name=$model_name"-wmt16-en-ro"
@@ -106,7 +106,7 @@ fi
 # need to switch to using latest version for model, currently blocked with a bug.
 # submit finetuning job
 parent_job_name=$( az ml job create --file ./translation-pipeline.yml $workspace_info --query name -o tsv --set \
-  jobs.translation_en_ro_finetune_job.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
+  jobs.translation_pipeline.component="azureml://registries/$registry_name/components/$finetuning_pipeline_component/labels/latest" \
   inputs.compute_model_import=$compute_cluster \
   inputs.compute_preprocess=$compute_cluster \
   inputs.compute_finetune=$compute_cluster \

sdk/python/foundation-models/system/finetune/question-answering/extractive-qa.ipynb

@@ -158,7 +158,7 @@
    "outputs": [],
    "source": [
     "model_name = \"bert-base-uncased\"\n",
-    "model_version = \"1\"\n",
+    "model_version = \"3\"\n",
     "foundation_model = registry_ml_client.models.get(model_name, model_version)\n",
     "print(\n",
     "    \"\\n\\nUsing model name: {0}, version: {1}, id: {2} for fine tuning\".format(\n",
@@ -250,7 +250,7 @@
     "# define the pipeline job\n",
     "@pipeline()\n",
     "def create_pipeline():\n",
-    "    finetuning_job = pipeline_component_func(\n",
+    "    question_answering_pipeline = pipeline_component_func(\n",
     "        # specify the foundation model available in the azureml system registry id identified in step #3\n",
     "        mlflow_model_path=foundation_model.id,\n",
     "        # huggingface_id = 'bert-base-uncased', # if you want to use a huggingface model, uncomment this line and comment the above line\n",
@@ -293,7 +293,7 @@
     "    return {\n",
     "        # map the output of the fine tuning job to the output of the pipeline job so that we can easily register the fine tuned model\n",
     "        # registering the model is required to deploy the model to an online or batch endpoint\n",
-    "        \"trained_model\": finetuning_job.outputs.mlflow_model_folder\n",
+    "        \"trained_model\": question_answering_pipeline.outputs.mlflow_model_folder\n",
     "    }\n",
     "\n",
     "\n",

sdk/python/foundation-models/system/finetune/summarization/news-summary.ipynb

@@ -156,7 +156,7 @@
    "outputs": [],
    "source": [
     "model_name = \"t5-small\"\n",
-    "model_version = \"1\"\n",
+    "model_version = \"3\"\n",
     "foundation_model = registry_ml_client.models.get(model_name, model_version)\n",
     "print(\n",
     "    \"\\n\\nUsing model name: {0}, version: {1}, id: {2} for fine tuning\".format(\n",
@@ -248,7 +248,7 @@
     "# define the pipeline job\n",
     "@pipeline()\n",
     "def create_pipeline():\n",
-    "    finetuning_job = pipeline_component_func(\n",
+    "    summarization_pipeline = pipeline_component_func(\n",
     "        # specify the foundation model available in the azureml system registry id identified in step #3\n",
     "        mlflow_model_path=foundation_model.id,\n",
     "        # huggingface_id = 'bert-base-uncased', # if you want to use a huggingface model, uncomment this line and comment the above line\n",
@@ -280,7 +280,7 @@
     "    return {\n",
     "        # map the output of the fine tuning job to the output of the pipeline job so that we can easily register the fine tuned model\n",
     "        # registering the model is required to deploy the model to an online or batch endpoint\n",
-    "        \"trained_model\": finetuning_job.outputs.mlflow_model_folder\n",
+    "        \"trained_model\": summarization_pipeline.outputs.mlflow_model_folder\n",
     "    }\n",
     "\n",
     "\n",

sdk/python/foundation-models/system/finetune/text-classification/emotion-detection.ipynb

@@ -157,7 +157,7 @@
    "outputs": [],
    "source": [
     "model_name = \"bert-base-uncased\"\n",
-    "model_version = \"1\"\n",
+    "model_version = \"3\"\n",
     "foundation_model = registry_ml_client.models.get(model_name, model_version)\n",
     "print(\n",
     "    \"\\n\\nUsing model name: {0}, version: {1}, id: {2} for fine tuning\".format(\n",
@@ -280,7 +280,7 @@
     "# define the pipeline job\n",
     "@pipeline()\n",
     "def create_pipeline():\n",
-    "    finetuning_job = pipeline_component_func(\n",
+    "    text_classification_pipeline = pipeline_component_func(\n",
     "        # specify the foundation model available in the azureml system registry id identified in step #3\n",
     "        mlflow_model_path=foundation_model.id,\n",
     "        # huggingface_id = 'bert-base-uncased', # if you want to use a huggingface model, uncomment this line and comment the above line\n",
@@ -298,9 +298,9 @@
     "        test_file_path=Input(\n",
     "            type=\"uri_file\", path=\"./emotion-dataset/small_test.jsonl\"\n",
     "        ),\n",
-    "        # evaluation_config=Input(\n",
-    "        #     type=\"uri_file\", path=\"./text-classification-config.json\"\n",
-    "        # ),\n",
+    "        evaluation_config=Input(\n",
+    "            type=\"uri_file\", path=\"./text-classification-config.json\"\n",
+    "        ),\n",
     "        # The following parameters map to the dataset fields\n",
     "        sentence1_key=\"text\",\n",
     "        label_key=\"label_string\",\n",
@@ -312,7 +312,7 @@
     "    return {\n",
     "        # map the output of the fine tuning job to the output of pipeline job so that we can easily register the fine tuned model\n",
     "        # registering the model is required to deploy the model to an online or batch endpoint\n",
-    "        \"trained_model\": finetuning_job.outputs.mlflow_model_folder\n",
+    "        \"trained_model\": text_classification_pipeline.outputs.mlflow_model_folder\n",
     "    }\n",
     "\n",
     "\n",

sdk/python/foundation-models/system/finetune/token-classification/token-classification.ipynb

@@ -157,7 +157,7 @@
    "outputs": [],
    "source": [
     "model_name = \"bert-base-uncased\"\n",
-    "model_version = \"1\"\n",
+    "model_version = \"3\"\n",
     "foundation_model = registry_ml_client.models.get(model_name, model_version)\n",
     "print(\n",
     "    \"\\n\\nUsing model name: {0}, version: {1}, id: {2} for fine tuning\".format(\n",
@@ -255,7 +255,7 @@
     "# define the pipeline job\n",
     "@pipeline()\n",
     "def create_pipeline():\n",
-    "    finetuning_job = pipeline_component_func(\n",
+    "    token_classification_pipeline = pipeline_component_func(\n",
     "        # specify the foundation model available in the azureml system registry id identified in step #3\n",
     "        mlflow_model_path=foundation_model.id,\n",
     "        # huggingface_id = 'bert-base-uncased', # if you want to use a huggingface model, uncomment this line and comment the above line\n",
@@ -287,7 +287,7 @@
     "    return {\n",
     "        # map the output of the fine tuning job to the output of pipeline job so that we can easily register the fine tuned model\n",
     "        # registering the model is required to deploy the model to an online or batch endpoint\n",
-    "        \"trained_model\": finetuning_job.outputs.mlflow_model_folder\n",
+    "        \"trained_model\": token_classification_pipeline.outputs.mlflow_model_folder\n",
     "    }\n",
     "\n",
     "\n",

sdk/python/foundation-models/system/finetune/translation/translation.ipynb

@@ -156,7 +156,7 @@
    "outputs": [],
    "source": [
     "model_name = \"t5-small\"\n",
-    "model_version = \"1\"\n",
+    "model_version = \"3\"\n",
     "foundation_model = registry_ml_client.models.get(model_name, model_version)\n",
     "print(\n",
     "    \"\\n\\nUsing model name: {0}, version: {1}, id: {2} for fine tuning\".format(\n",
@@ -245,7 +245,7 @@
     "# define the pipeline job\n",
     "@pipeline()\n",
     "def create_pipeline():\n",
-    "    finetuning_job = pipeline_component_func(\n",
+    "    translation_pipeline = pipeline_component_func(\n",
     "        # specify the foundation model available in the azureml system registry id identified in step #3\n",
     "        mlflow_model_path=foundation_model.id,\n",
     "        # huggingface_id = 't5-small', # if you want to use a huggingface model, uncomment this line and comment the above line\n",
@@ -277,7 +277,7 @@
     "    return {\n",
     "        # map the output of the fine tuning job to the output of the pipeline job so that we can easily register the fine tuned model\n",
     "        # registering the model is required to deploy the model to an online or batch endpoint\n",
-    "        \"trained_model\": finetuning_job.outputs.mlflow_model_folder\n",
+    "        \"trained_model\": translation_pipeline.outputs.mlflow_model_folder\n",
     "    }\n",
     "\n",
     "\n",