modelParameters.html

<div id="tabby_loader_popup">
    <div>
        <h3><strong data-i18n="">Set Parameters</strong>
            <a href="https://github.com/theroyallab/tabbyAPI" class="notes-link" target="_blank">
                <span class="note-link-span">?</span>
            </a>
        </h3>

        <small class="flex-container extensions_info justifyCenter">
            Set Parameters for Loading a Model
        </small>
        <hr />

        <!-- Model parameter textboxes -->
        Main Model
        <div class="flex-container">
            <div class="flex1">
                <label for="max_seq_len">
                    <small data-i18n="Max Seq Len">Max Seq Len</small>
                </label>
                <input name="max_seq_len" class="text_pole" type="text" />
            </div>
            <div class="flex1">
                <label for="rope_scale">
                    <small data-i18n="Rope Scale">Rope Scale</small>
                </label>
                <input name="rope_scale" class="text_pole" type="text" />
            </div>
            <div class="flex1">
                <label for="rope_alpha">
                    <small data-i18n="Rope Alpha">Rope Alpha</small>
                </label>
                <input name="rope_alpha" class="text_pole" type="text" />
            </div>
        </div>
        Draft Model
        <div class="flex-container">
            <div class="flex1">
                <label for="rope_scale">
                    <small data-i18n="Rope Scale">Rope Scale</small>
                </label>
                <input name="draft_rope_scale" class="text_pole" type="text" />
            </div>
            <div class="flex1">
                <label for="rope_alpha">
                    <small data-i18n="Rope Alpha">Rope Alpha</small>
                </label>
                <input name="draft_rope_alpha" class="text_pole" type="text" />
            </div>
        </div>

        <!-- Container for GPU Split and other options -->
        Loader Options
        <div class="flex-container padded-loader-setting-category">
            <div class="flex-container flexFlowColumn flexNoGap alignitemsstart">
                <small class="justifyCenter">GPU Split</small>
                <label class="checkbox flex-container">
                    <input type="checkbox" name="gpu_split_auto" />
                    <span data-i18n="Auto Split">Auto Split</span>
                </label>
                <input name="gpu_split_value" class="text_pole" type="text" placeholder="ex. [20.6, 24]" />
            </div>

            <div class="flex-container flexFlowColumn flexNoGap alignitemsstart padded-loader-setting-block">
                <small>Other Options</small>
                <label class="checkbox flex-container">
                    <input type="checkbox" name="no_flash_attention" />
                    <span data-i18n="Disable FA2">Disable FA2</span>
                </label>
                <label class="checkbox flex-container">
                    <input type="checkbox" name="use_cfg" />
                    <span data-i18n="Use CFG">Use CFG</span>
                </label>
                <label class="checkbox flex-container">
                    <input type="checkbox" name="fasttensors" />
                    <span data-i18n="FastTensors">FastTensors</span>
                </label>
            </div>

            <div class="flex-container flexFlowColumn flexNoGap alignitemsstart">
                <small class="justifyCenter">Cache Mode</small>
                <select name="cache_mode_select" class="margin0">
                    <option value="0">FP16</option>
                    <option value="1">Q4</option>
                    <option value="2">Q6</option>
                    <option value="3">Q8</option>
                </select>
            </div>
        </div>
    </div>
</div>