Default name updates

python/sparknlp/annotator/embeddings/nomic_embeddings.py

@@ -31,7 +31,7 @@ class NomicEmbeddings(AnnotatorModel, HasEmbeddingsProperties, HasCaseSensitiveP
     ...     .setOutputCol("nomic_embeddings")
 
 
-    The default model is ``"nomic_small"``, if no name is provided.
+    The default model is ``"nomic_embed_v1"``, if no name is provided.
 
     For available pretrained models please see the
     `Models Hub <https://sparknlp.org/models?q=Nomic>`__.
@@ -159,13 +159,13 @@ def loadSavedModel(folder, spark_session, use_openvino=False):
         return NomicEmbeddings(java_model=jModel)
 
     @staticmethod
-    def pretrained(name="nomic_small", lang="en", remote_loc=None):
+    def pretrained(name="nomic_embed_v1", lang="en", remote_loc=None):
         """Downloads and loads a pretrained model.
 
         Parameters
         ----------
         name : str, optional
-            Name of the pretrained model, by default "nomic_small"
+            Name of the pretrained model, by default "nomic_embed_v1"
         lang : str, optional
             Language of the pretrained model, by default "en"
         remote_loc : str, optional

python/sparknlp/annotator/seq2seq/cpm_transformer.py

@@ -44,7 +44,7 @@ class CPMTransformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     ...     .setOutputCol("generation")
 
 
-    The default model is ``"llam2-7b"``, if no name is provided. For available
+    The default model is ``"mini_cpm_2b_8bit"``, if no name is provided. For available
     pretrained models please see the `Models Hub
     <https://sparknlp.org/models?q=cpm>`__.
 
@@ -104,7 +104,7 @@ class CPMTransformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     >>> documentAssembler = DocumentAssembler() \\
     ...     .setInputCol("text") \\
     ...     .setOutputCol("documents")
-    >>> cpm = CPMTransformer.pretrained("llama_2_7b_chat_hf_int4") \\
+    >>> cpm = CPMTransformer.pretrained("mini_cpm_2b_8bit","xx") \\
     ...     .setInputCols(["documents"]) \\
     ...     .setMaxOutputLength(50) \\
     ...     .setOutputCol("generation")
@@ -299,15 +299,15 @@ def loadSavedModel(folder, spark_session,  use_openvino = False):
         return CPMTransformer(java_model=jModel)
 
     @staticmethod
-    def pretrained(name="llama_2_7b_chat_hf_int4", lang="en", remote_loc=None):
+    def pretrained(name="mini_cpm_2b_8bit", lang="xx", remote_loc=None):
         """Downloads and loads a pretrained model.
 
         Parameters
         ----------
         name : str, optional
-            Name of the pretrained model, by default "llama_2_7b_chat_hf_int4"
+            Name of the pretrained model, by default "mini_cpm_2b_8bit"
         lang : str, optional
-            Language of the pretrained model, by default "en"
+            Language of the pretrained model, by default "xx"
         remote_loc : str, optional
             Optional remote address of the resource, by default None. Will use
             Spark NLPs repositories otherwise.

python/sparknlp/annotator/seq2seq/nllb_transformer.py

@@ -32,7 +32,7 @@ class NLLBTransformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     ...     .setOutputCol("generation")
 
 
-    The default model is ``"nllb_418M"``, if no name is provided. For available
+    The default model is ``"nllb_distilled_600M_8int"``, if no name is provided. For available
     pretrained models please see the `Models Hub
     <https://sparknlp.org/models?q=nllb>`__.
 
@@ -164,7 +164,7 @@ class NLLBTransformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     >>> documentAssembler = DocumentAssembler() \\
     ...     .setInputCol("text") \\
     ...     .setOutputCol("documents")
-    >>> nllb = NLLBTransformer.pretrained("nllb_418M") \\
+    >>> nllb = NLLBTransformer.pretrained("nllb_distilled_600M_8int") \\
     ...     .setInputCols(["documents"]) \\
     ...     .setMaxOutputLength(50) \\
     ...     .setOutputCol("generation") \\
@@ -398,13 +398,13 @@ def loadSavedModel(folder, spark_session,  use_openvino=False):
         return NLLBTransformer(java_model=jModel)
 
     @staticmethod
-    def pretrained(name="nllb_418M", lang="xx", remote_loc=None):
+    def pretrained(name="nllb_distilled_600M_8int", lang="xx", remote_loc=None):
         """Downloads and loads a pretrained model.
 
         Parameters
         ----------
         name : str, optional
-            Name of the pretrained model, by default "nllb_418M"
+            Name of the pretrained model, by default "nllb_distilled_600M_8int"
         lang : str, optional
             Language of the pretrained model, by default "en"
         remote_loc : str, optional

python/sparknlp/annotator/seq2seq/phi3_transformer.py

@@ -37,7 +37,7 @@ class Phi3Transformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     ...     .setOutputCol("generation")
 
 
-    The default model is ``"phi3"``, if no name is provided. For available
+    The default model is ``phi_3_mini_128k_instruct``, if no name is provided. For available
     pretrained models please see the `Models Hub
     <https://sparknlp.org/models?q=phi3>`__.
 
@@ -112,7 +112,7 @@ class Phi3Transformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     >>> documentAssembler = DocumentAssembler() \\
     ...     .setInputCol("text") \\
     ...     .setOutputCol("documents")
-    >>> phi3 = Phi3Transformer.pretrained("phi3") \\
+    >>> phi3 = Phi3Transformer.pretrained(phi_3_mini_128k_instruct) \\
     ...     .setInputCols(["documents"]) \\
     ...     .setMaxOutputLength(50) \\
     ...     .setOutputCol("generation")
@@ -308,13 +308,13 @@ def loadSavedModel(folder, spark_session, use_openvino=False):
         return Phi3Transformer(java_model=jModel)
 
     @staticmethod
-    def pretrained(name="phi3", lang="en", remote_loc=None):
+    def pretrained(name="phi_3_mini_128k_instruct", lang="en", remote_loc=None):
         """Downloads and loads a pretrained model.
 
         Parameters
         ----------
         name : str, optional
-            Name of the pretrained model, by default "phi3"
+            Name of the pretrained model, by default phi_3_mini_128k_instruct
         lang : str, optional
             Language of the pretrained model, by default "en"
         remote_loc : str, optional

python/sparknlp/annotator/seq2seq/qwen_transformer.py

@@ -121,7 +121,7 @@ class QwenTransformer(AnnotatorModel, HasBatchedAnnotate, HasEngine):
     >>> documentAssembler = DocumentAssembler() \\
     ...     .setInputCol("text") \\
     ...     .setOutputCol("documents")
-    >>> qwen = QwenTransformer.pretrained("qwen-7b") \\
+    >>> qwen = QwenTransformer.pretrained("qwen_7.5b_chat") \\
     ...     .setInputCols(["documents"]) \\
     ...     .setMaxOutputLength(50) \\
     ...     .setOutputCol("generation")
@@ -317,13 +317,13 @@ def loadSavedModel(folder, spark_session, use_openvino=False):
         return QwenTransformer(java_model=jModel)
 
     @staticmethod
-    def pretrained(name="qwen-7b", lang="en", remote_loc=None):
+    def pretrained(name="qwen_7.5b_chat", lang="en", remote_loc=None):
         """Downloads and loads a pretrained model.
 
         Parameters
         ----------
         name : str, optional
-            Name of the pretrained model, by default "qwen-7b"
+            Name of the pretrained model, by default "qwen_7.5b_chat"
         lang : str, optional
             Language of the pretrained model, by default "en"
         remote_loc : str, optional

src/main/scala/com/johnsnowlabs/nlp/annotators/seq2seq/CPMTransformer.scala

@@ -68,7 +68,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCols("document")
   *   .setOutputCol("generation")
   * }}}
-  * The default model is `"llama_2_7b_chat_hf_int4"`, if no name is provided. For available
+  * The default model is `"mini_cpm_2b_8bit"`, if no name is provided. For available
   * pretrained models please see the [[https://sparknlp.org/models?q=cpm Models Hub]].
   *
   * For extended examples of usage, see
@@ -94,7 +94,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCol("text")
   *   .setOutputCol("documents")
   *
-  * val cpm = CPMTransformer.pretrained("llama_2_7b_chat_hf_int4")
+  * val cpm = CPMTransformer.pretrained("mini_cpm_2b_8bit")
   *   .setInputCols(Array("documents"))
   *   .setMinOutputLength(10)
   *   .setMaxOutputLength(50)
@@ -311,7 +311,8 @@ class CPMTransformer(override val uid: String)
 trait ReadablePretrainedCPMTransformerModel
     extends ParamsAndFeaturesReadable[CPMTransformer]
     with HasPretrained[CPMTransformer] {
-  override val defaultModelName: Some[String] = Some("llama_2_7b_chat_hf_int4")
+  override val defaultModelName: Some[String] = Some("mini_cpm_2b_8bit")
+  override val defaultLang: String = "xx"
 
   /** Java compliant-overrides */
   override def pretrained(): CPMTransformer = super.pretrained()

src/main/scala/com/johnsnowlabs/nlp/annotators/seq2seq/NLLBTransformer.scala

@@ -59,7 +59,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCols("document")
   *   .setOutputCol("generation")
   * }}}
-  * The default model is `"nllb_418M"`, if no name is provided. For available pretrained models
+  * The default model is `"nllb_distilled_600M_8int"`, if no name is provided. For available pretrained models
   * please see the [[https://sparknlp.org/models?q=nllb Models Hub]].
   *
   * For extended examples of usage, see
@@ -156,7 +156,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCol("text")
   *   .setOutputCol("documents")
   *
-  * val nllb = NLLBTransformer.pretrained("nllb_418M")
+  * val nllb = NLLBTransformer.pretrained("nllb_distilled_600M_8int")
   *   .setInputCols(Array("documents"))
   *   .setSrcLang("zho_Hans")
   *   .serTgtLang("eng_Latn")
@@ -635,7 +635,7 @@ class NLLBTransformer(override val uid: String)
 trait ReadablePretrainedNLLBTransformerModel
     extends ParamsAndFeaturesReadable[NLLBTransformer]
     with HasPretrained[NLLBTransformer] {
-  override val defaultModelName: Some[String] = Some("nllb_418M")
+  override val defaultModelName: Some[String] = Some("nllb_distilled_600M_8int")
   override val defaultLang: String = "xx"
 
   /** Java compliant-overrides */

src/main/scala/com/johnsnowlabs/nlp/annotators/seq2seq/Phi3Transformer.scala

@@ -65,7 +65,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCols("document")
   *   .setOutputCol("generation")
   * }}}
-  * The default model is `"phi_3_mini_128k_instruct_int8"`, if no name is provided. For available
+  * The default model is `"phi_3_mini_128k_instruct"`, if no name is provided. For available
   * pretrained models please see the [[https://sparknlp.org/models?q=phi3 Models Hub]].
   *
   * For extended examples of usage, see
@@ -106,7 +106,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCol("text")
   *   .setOutputCol("documents")
   *
-  * val phi3 = Phi3Transformer.pretrained("phi_3_mini_128k_instruct_int8")
+  * val phi3 = Phi3Transformer.pretrained("phi_3_mini_128k_instruct")
   *   .setInputCols(Array("documents"))
   *   .setMinOutputLength(10)
   *   .setMaxOutputLength(50)
@@ -323,7 +323,7 @@ class Phi3Transformer(override val uid: String)
 trait ReadablePretrainedPhi3TransformerModel
     extends ParamsAndFeaturesReadable[Phi3Transformer]
     with HasPretrained[Phi3Transformer] {
-  override val defaultModelName: Some[String] = Some("phi_3_mini_128k_instruct_int8")
+  override val defaultModelName: Some[String] = Some("phi_3_mini_128k_instruct")
 
   /** Java compliant-overrides */
   override def pretrained(): Phi3Transformer = super.pretrained()

src/main/scala/com/johnsnowlabs/nlp/annotators/seq2seq/QwenTransformer.scala

@@ -68,7 +68,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCols("document")
   *   .setOutputCol("generation")
   * }}}
-  * The default model is `"Qwen-13b"`, if no name is provided. For available pretrained models
+  * The default model is `"qwen_7.5b_chat"`, if no name is provided. For available pretrained models
   * please see the [[https://sparknlp.org/models?q=Qwen Models Hub]].
   *
   * For extended examples of usage, see
@@ -113,7 +113,7 @@ import org.json4s.jackson.JsonMethods._
   *   .setInputCol("text")
   *   .setOutputCol("documents")
   *
-  * val Qwen = QwenTransformer.pretrained("Qwen-7b")
+  * val Qwen = QwenTransformer.pretrained("qwen_7.5b_chat")
   *   .setInputCols(Array("documents"))
   *   .setMinOutputLength(10)
   *   .setMaxOutputLength(50)
@@ -334,7 +334,7 @@ class QwenTransformer(override val uid: String)
 trait ReadablePretrainedQwenTransformerModel
     extends ParamsAndFeaturesReadable[QwenTransformer]
     with HasPretrained[QwenTransformer] {
-  override val defaultModelName: Some[String] = Some("Qwen-7b")
+  override val defaultModelName: Some[String] = Some("qwen_7.5b_chat")
 
   /** Java compliant-overrides */
   override def pretrained(): QwenTransformer = super.pretrained()

src/main/scala/com/johnsnowlabs/nlp/embeddings/NomicEmbeddings.scala

@@ -49,7 +49,7 @@ import com.johnsnowlabs.ml.openvino.{OpenvinoWrapper, ReadOpenvinoModel, WriteOp
   *   .setInputCols("document")
   *   .setOutputCol("nomic_embeddings")
   * }}}
-  * The default model is `"nomic_small"`, if no name is provided.
+  * The default model is `"nomic_embed_v1"`, if no name is provided.
   *
   * For available pretrained models please see the
   * [[https://sparknlp.org/models?q=NomicEmbeddings Models Hub]].
@@ -86,7 +86,7 @@ import com.johnsnowlabs.ml.openvino.{OpenvinoWrapper, ReadOpenvinoModel, WriteOp
   *   .setInputCol("text")
   *   .setOutputCol("document")
   *
-  * val embeddings = NomicEmbeddings.pretrained("nomic_small", "en")
+  * val embeddings = NomicEmbeddings.pretrained("nomic_embed_v1", "en")
   *   .setInputCols("document")
   *   .setOutputCol("nomic_embeddings")
   *
@@ -357,7 +357,7 @@ class NomicEmbeddings(override val uid: String)
 trait ReadablePretrainedNomicEmbeddingsModel
     extends ParamsAndFeaturesReadable[NomicEmbeddings]
     with HasPretrained[NomicEmbeddings] {
-  override val defaultModelName: Some[String] = Some("nomic_small")
+  override val defaultModelName: Some[String] = Some("nomic_embed_v1")
 
   /** Java compliant-overrides */
   override def pretrained(): NomicEmbeddings = super.pretrained()

src/main/scala/com/johnsnowlabs/nlp/pretrained/ResourceDownloader.scala

@@ -690,7 +690,12 @@ object PythonResourceDownloader {
     "SnowFlakeEmbeddings" -> SnowFlakeEmbeddings,
     "CamemBertForZeroShotClassification" -> CamemBertForZeroShotClassification,
     "BertForMultipleChoice" -> BertForMultipleChoice,
-    "PromptAssembler" -> PromptAssembler)
+    "PromptAssembler" -> PromptAssembler,
+    "CPMTransformer"-> CPMTransformer,
+    "NomicEmbeddings" -> NomicEmbeddings,
+    "NLLBTransformer" -> NLLBTransformer,
+    "Phi3Transformer" -> Phi3Transformer,
+    "QwenTransformer" -> QwenTransformer)
 
   // List pairs of types such as the one with key type can load a pretrained model from the value type
   val typeMapper: Map[String, String] = Map("ZeroShotNerModel" -> "RoBertaForQuestionAnswering")