Fix OpenLLaMA model names (#209)

* Update PR template. (#207)

* Update OpenLLaMA model names.

Author: rmitsch

.github/PULL_REQUEST_TEMPLATE.md

@@ -7,6 +7,9 @@ ran. If your test fixes a bug reported in an issue, don't forget to include the
 issue number. If your PR is still a work in progress, that's totally fine – just
 include a note to let us know. -->
 
+### Corresponding documentation PR
+<!--- Add the link to the corresponding documentation PR here, if applicable. -->
+
 ### Types of change
 <!-- What type of change does your PR cover? Is it a bug fix, an enhancement
 or new feature, or a change to the documentation? -->

README.md

@@ -1475,14 +1475,14 @@ Example config block:
 ```ini
 [components.llm.model]
 @llm_models = "spacy.OpenLLaMA.v1"
-name = "open_llama_3b_350bt_preview"
+name = "open_llama_3b"
 ```
 
-| Argument      | Type                                                                                                                                  | Default | Description                                                                                                                  |
-| ------------- | ------------------------------------------------------------------------------------------------------------------------------------- | ------- | ---------------------------------------------------------------------------------------------------------------------------- |
-| `name`        | `Literal["open_llama_3b_350bt_preview", "open_llama_3b_600bt_preview", "open_llama_7b_400bt_preview", "open_llama_7b_600bt_preview"]` |         | The name of a OpenLLaMA model that is supported (e. g. "open_llama_3b_350bt_preview").                                       |
-| `config_init` | `Dict[str, Any]`                                                                                                                      | `{}`    | Further configuration passed on to the construction of the model with `transformers.AutoModelForCausalLM.from_pretrained()`. |
-| `config_run`  | `Dict[str, Any]`                                                                                                                      | `{}`    | Further configuration used during model inference.                                                                           |
+| Argument      | Type                                                                                            | Default | Description                                                                                                                  |
+| ------------- |-------------------------------------------------------------------------------------------------| ------- | ---------------------------------------------------------------------------------------------------------------------------- |
+| `name`        | `Literal["open_llama_3b", "open_llama_7b", "open_llama_7b_v2", "open_llama_13b"]` |         | The name of a OpenLLaMA model that is supported.                                       |
+| `config_init` | `Dict[str, Any]`                                                                                | `{}`    | Further configuration passed on to the construction of the model with `transformers.AutoModelForCausalLM.from_pretrained()`. |
+| `config_run`  | `Dict[str, Any]`                                                                                | `{}`    | Further configuration used during model inference.                                                                           |
 
 See the [OpenLM Research OpenLLaMA GitHub repo](https://github.com/openlm-research/open_llama) for details.
 

spacy_llm/models/hf/openllama.py

@@ -9,10 +9,10 @@
 
 class OpenLLaMA(HuggingFace):
     MODEL_NAMES = Literal[
-        "open_llama_3b_350bt_preview",  # noqa: F722
-        "open_llama_3b_600bt_preview",  # noqa: F722
-        "open_llama_7b_400bt_preview",  # noqa: F722
-        "open_llama_7b_600bt_preview",  # noqa: F722
+        "open_llama_3b",  # noqa: F722
+        "open_llama_7b",  # noqa: F722
+        "open_llama_7b_v2",  # noqa: F722
+        "open_llama_13b",  # noqa: F722
     ]
 
     def __init__(

spacy_llm/tests/models/test_openllama.py

@@ -8,7 +8,7 @@
 _PIPE_CFG = {
     "model": {
         "@llm_models": "spacy.OpenLLaMA.v1",
-        "name": "open_llama_3b_350bt_preview",
+        "name": "open_llama_3b",
     },
     "task": {"@llm_tasks": "spacy.NoOp.v1"},
 }
@@ -29,7 +29,7 @@
 
 [components.llm.model]
 @llm_models = spacy.OpenLLaMA.v1
-name = open_llama_3b_350bt_preview
+name = open_llama_3b
 """