Merge pull request #5220 from oobabooga/dev

Merge dev branch

css/main.css

@@ -195,8 +195,9 @@ button {
     border: var(--input-border-width) solid var(--input-border-color) !important;
 }
 
-.file-saver > :first-child > :nth-child(2) {
+.file-saver > :first-child > :last-child {
     background: var(--block-background-fill);
+    justify-content: flex-end;
 }
 
 .checkboxgroup-table label {
@@ -652,14 +653,24 @@ div.svelte-362y77>*, div.svelte-362y77>.form>* {
 /* ----------------------------------------------
   Past chats menus
 ---------------------------------------------- */
-#past-chats-row {
-    margin-bottom: calc( -1 * var(--layout-gap) );
-}
-
 #rename-row label {
     margin-top: var(--layout-gap);
 }
 
+/* ----------------------------------------------
+  Past chat histories in a side bar on desktop
+---------------------------------------------- */
+@media screen and (width >= 1327px) {
+    #past-chats-row {
+        position: absolute;
+        top: 16px;
+        left: 0;
+        width: calc(0.5*(var(--document-width) - 880px - 120px - 16px*2));
+        max-width: 300px;
+        margin-left: calc(-0.5*(var(--document-width) - 880px - 14px - 16px * 2));
+    }
+}
+
 /* ----------------------------------------------
   Keep dropdown menus above errored components
 ---------------------------------------------- */

docs/03 - Parameters Tab.md

@@ -21,7 +21,7 @@ These were obtained after a blind contest called "Preset Arena" where hundreds o
 
 A key takeaway is that the best presets are:
 
-* **For Instruct**: Divine Intellect, Big O, simple-1, Space Alien, StarChat, Titanic, tfs-with-top-a, Asterism, Contrastive Search (only works for the Transformers loader at the moment).
+* **For Instruct**: Divine Intellect, Big O, simple-1.
 * **For Chat**: Midnight Enigma, Yara, Shortwave.
 
 The other presets are:
@@ -54,8 +54,7 @@ For more information about the parameters, the [transformers documentation](http
 * **mirostat_mode**: Activates the Mirostat sampling technique. It aims to control perplexity during sampling. See the [paper](https://arxiv.org/abs/2007.14966).
 * **mirostat_tau**: No idea, see the paper for details. According to the Preset Arena, 8 is a good value. 
 * **mirostat_eta**: No idea, see the paper for details. According to the Preset Arena, 0.1 is a good value.
-* **dynamic_temperature_low**: The lower bound for temperature in Dynamic Temperature. Only used when "dynamic_temperature" is checked.
-* **dynamic_temperature**: Activates Dynamic Temperature. This modifies temperature to range between "dynamic_temperature_low" (minimum) and "temperature" (maximum), with an entropy-based scaling.
+* **dynamic_temperature**: Activates Dynamic Temperature. This modifies temperature to range between "dynatemp_low" (minimum) and "dynatemp_high" (maximum), with an entropy-based scaling. The steepness of the curve is controlled by "dynatemp_exponent".
 * **temperature_last**: Makes temperature the last sampler instead of the first. With this, you can remove low probability tokens with a sampler like min_p and then use a high temperature to make the model creative without losing coherency.
 * **do_sample**: When unchecked, sampling is entirely disabled, and greedy decoding is used instead (the most likely token is always picked).
 * **Seed**: Set the Pytorch seed to this number. Note that some loaders do not use Pytorch (notably llama.cpp), and others are not deterministic (notably ExLlama v1 and v2). For these loaders, the seed has no effect.

extensions/gallery/script.py

@@ -108,7 +108,7 @@ def ui():
         gr.HTML(value="<style>" + generate_css() + "</style>")
         with gr.Row():
             filter_box = gr.Textbox(label='', placeholder='Filter', lines=1, max_lines=1, container=False, elem_id='gallery-filter-box')
-            gr.ClearButton(filter_box, value='🗑️', elem_classes='refresh-button')
+            gr.ClearButton(filter_box, value='Clear', elem_classes='refresh-button')
             update = gr.Button("Refresh", elem_classes='refresh-button')
 
         gallery = gr.Dataset(

extensions/openai/typing.py

@@ -9,7 +9,9 @@ class GenerationOptions(BaseModel):
     preset: str | None = Field(default=None, description="The name of a file under text-generation-webui/presets (without the .yaml extension). The sampling parameters that get overwritten by this option are the keys in the default_preset() function in modules/presets.py.")
     min_p: float = 0
     dynamic_temperature: bool = False
-    dynamic_temperature_low: float = 0.1
+    dynatemp_low: float = 1
+    dynatemp_high: float = 1
+    dynatemp_exponent: float = 1
     top_k: int = 0
     repetition_penalty: float = 1
     repetition_penalty_range: int = 1024

instruction-templates/Mistral.yaml

@@ -4,13 +4,12 @@ instruction_template: |-
           {{- message['content'] -}}
       {%- else -%}
           {%- if message['role'] == 'user' -%}
-              {{-'[INST] ' + message['content'] + ' [/INST]'-}}
+              {{-' [INST] ' + message['content'].rstrip() + ' [/INST] '-}}
           {%- else -%}
-              {{-'' + message['content'] + '</s> ' -}}
+              {{-'' + message['content'] + '</s>' -}}
           {%- endif -%}
       {%- endif -%}
   {%- endfor -%}
   {%- if add_generation_prompt -%}
       {{-''-}}
-  {%- endif -%}
-
+  {%- endif -%}

js/main.js

@@ -37,6 +37,7 @@ document.querySelector(".header_bar").addEventListener("click", function(event)
 //------------------------------------------------
 // Keyboard shortcuts
 //------------------------------------------------
+let previousTabId = "chat-tab-button";
 document.addEventListener("keydown", function(event) {
 
   // Stop generation on Esc pressed
@@ -97,6 +98,19 @@ document.addEventListener("keydown", function(event) {
     document.getElementById("Impersonate").click();
   }
 
+  // Switch between tabs on Tab
+  else if (!event.ctrlKey && !event.shiftKey && event.key === "Tab") {
+    event.preventDefault();
+    var parametersButton = document.getElementById("parameters-button");
+    var parentContainer = parametersButton.parentNode;
+    var selectedChild = parentContainer.querySelector(".selected");
+    if (selectedChild.id == "parameters-button") {
+      document.getElementById(previousTabId).click();
+    } else {
+      previousTabId = selectedChild.id;
+      parametersButton.click();
+    }
+  }
 });
 
 //------------------------------------------------
@@ -248,7 +262,7 @@ button.addEventListener("click", function () {
     hideMenu();
   }
   else {
-  showMenu();
+    showMenu();
   }
 });
 
@@ -371,3 +385,43 @@ new ResizeObserver(updateCssProperties)
   .observe(document.querySelector("#chat-input textarea"));
 
 window.addEventListener("resize", updateCssProperties);
+
+//------------------------------------------------
+// Keep track of the display width to position the past
+// chats dropdown on desktop
+//------------------------------------------------
+function updateDocumentWidth() {
+  var updatedWidth = window.innerWidth || document.documentElement.clientWidth || document.body.clientWidth;
+  document.documentElement.style.setProperty("--document-width", updatedWidth + "px");
+}
+
+updateDocumentWidth();
+window.addEventListener("resize", updateDocumentWidth);
+
+//------------------------------------------------
+// Focus on the rename text area when it becomes visible
+//------------------------------------------------
+const renameTextArea = document.getElementById("rename-row").querySelector("textarea");
+
+function respondToVisibility(element, callback) {
+  var options = {
+    root: document.documentElement,
+  };
+
+  var observer = new IntersectionObserver((entries, observer) => {
+    entries.forEach(entry => {
+      callback(entry.intersectionRatio > 0);
+    });
+  }, options);
+
+  observer.observe(element);
+}
+
+
+function handleVisibilityChange(isVisible) {
+  if (isVisible) {
+    renameTextArea.focus();
+  }
+}
+
+respondToVisibility(renameTextArea, handleVisibilityChange);

modules/RoPE.py

@@ -3,7 +3,7 @@ def get_alpha_value(alpha, base):
     Gets alpha_value from alpha_value and rope_freq_base
     '''
     if base > 0:
-        return (base/10000.) ** (63/64.)
+        return (base / 10000.) ** (63 / 64.)
     else:
         return alpha
 
@@ -15,4 +15,4 @@ def get_rope_freq_base(alpha, base):
     if base > 0:
         return base
     else:
-        return 10000 * alpha ** (64/63.)
+        return 10000 * alpha ** (64 / 63.)

modules/chat.py

@@ -14,6 +14,7 @@
 from PIL import Image
 
 import modules.shared as shared
+from modules import utils
 from modules.extensions import apply_extensions
 from modules.html_generator import chat_html_wrapper, make_thumbnail
 from modules.logging_colors import logger
@@ -517,6 +518,35 @@ def load_latest_history(state):
     return history
 
 
+def load_history_after_deletion(state, idx):
+    '''
+    Loads the latest history for the given character in chat or chat-instruct
+    mode, or the latest instruct history for instruct mode.
+    '''
+
+    if shared.args.multi_user:
+        return start_new_chat(state)
+
+    histories = find_all_histories(state)
+    idx = min(int(idx), len(histories) - 1)
+    idx = max(0, idx)
+
+    if len(histories) > 0:
+        history = load_history(histories[idx], state['character_menu'], state['mode'])
+    else:
+        history = start_new_chat(state)
+        histories = find_all_histories(state)
+
+    return history, gr.update(choices=histories, value=histories[idx])
+
+
+def update_character_menu_after_deletion(idx):
+    characters = utils.get_available_characters()
+    idx = min(int(idx), len(characters) - 1)
+    idx = max(0, idx)
+    return gr.update(choices=characters, value=characters[idx])
+
+
 def load_history(unique_id, character, mode):
     p = get_history_file_path(unique_id, character, mode)
 

modules/loaders.py

@@ -156,7 +156,9 @@ def transformers_samplers():
         'temperature',
         'temperature_last',
         'dynamic_temperature',
-        'dynamic_temperature_low',
+        'dynatemp_low',
+        'dynatemp_high',
+        'dynatemp_exponent',
         'top_p',
         'min_p',
         'top_k',
@@ -223,7 +225,9 @@ def transformers_samplers():
         'temperature',
         'temperature_last',
         'dynamic_temperature',
-        'dynamic_temperature_low',
+        'dynatemp_low',
+        'dynatemp_high',
+        'dynatemp_exponent',
         'top_p',
         'min_p',
         'top_k',
@@ -277,7 +281,9 @@ def transformers_samplers():
         'temperature',
         'temperature_last',
         'dynamic_temperature',
-        'dynamic_temperature_low',
+        'dynatemp_low',
+        'dynatemp_high',
+        'dynatemp_exponent',
         'top_p',
         'min_p',
         'top_k',
@@ -350,12 +356,20 @@ def list_all_samplers():
     return sorted(all_samplers)
 
 
-def blacklist_samplers(loader):
+def blacklist_samplers(loader, dynamic_temperature):
     all_samplers = list_all_samplers()
-    if loader == 'All':
-        return [gr.update(visible=True) for sampler in all_samplers]
-    else:
-        return [gr.update(visible=True) if sampler in loaders_samplers[loader] else gr.update(visible=False) for sampler in all_samplers]
+    output = []
+
+    for sampler in all_samplers:
+        if loader == 'All' or sampler in loaders_samplers[loader]:
+            if sampler.startswith('dynatemp'):
+                output.append(gr.update(visible=dynamic_temperature))
+            else:
+                output.append(gr.update(visible=True))
+        else:
+            output.append(gr.update(visible=False))
+
+    return output
 
 
 def get_model_types(loader):

modules/models.py

@@ -309,14 +309,14 @@ def AutoAWQ_loader(model_name):
     model_dir = Path(f'{shared.args.model_dir}/{model_name}')
 
     model = AutoAWQForCausalLM.from_quantized(
-                quant_path=model_dir,
-                max_new_tokens=shared.args.max_seq_len,
-                trust_remote_code=shared.args.trust_remote_code,
-                fuse_layers=not shared.args.no_inject_fused_attention,
-                max_memory=get_max_memory_dict(),
-                batch_size=1,
-                safetensors=any(model_dir.glob('*.safetensors')),
-            )
+        quant_path=model_dir,
+        max_new_tokens=shared.args.max_seq_len,
+        trust_remote_code=shared.args.trust_remote_code,
+        fuse_layers=not shared.args.no_inject_fused_attention,
+        max_memory=get_max_memory_dict(),
+        batch_size=1,
+        safetensors=any(model_dir.glob('*.safetensors')),
+    )
 
     return model
 

modules/presets.py

@@ -6,14 +6,17 @@
 
 from modules import shared
 from modules.loaders import loaders_samplers
+from modules.logging_colors import logger
 
 
 def default_preset():
     return {
         'temperature': 1,
         'temperature_last': False,
         'dynamic_temperature': False,
-        'dynamic_temperature_low': 0.1,
+        'dynatemp_low': 1,
+        'dynatemp_high': 1,
+        'dynatemp_exponent': 1,
         'top_p': 1,
         'min_p': 0,
         'top_k': 0,
@@ -48,11 +51,15 @@ def presets_params():
 def load_preset(name):
     generate_params = default_preset()
     if name not in ['None', None, '']:
-        with open(Path(f'presets/{name}.yaml'), 'r') as infile:
-            preset = yaml.safe_load(infile)
-
-        for k in preset:
-            generate_params[k] = preset[k]
+        path = Path(f'presets/{name}.yaml')
+        if path.exists():
+            with open(path, 'r') as infile:
+                preset = yaml.safe_load(infile)
+
+            for k in preset:
+                generate_params[k] = preset[k]
+        else:
+            logger.error(f"The preset \"{name}\" does not exist under \"{path}\". Using the default parameters.")
 
     return generate_params