mirror of
https://github.com/labmlai/annotated_deep_learning_paper_implementations.git
synced 2025-08-14 17:41:37 +08:00
LoRA docs
This commit is contained in:
Varuna Jayasiri
@ -101,6 +101,7 @@
|
||||
<li><a href="transformers/vit/index.html">Vision Transformer (ViT)</a> </li>
|
||||
<li><a href="transformers/primer_ez/index.html">Primer EZ</a> </li>
|
||||
<li><a href="transformers/hour_glass/index.html">Hourglass</a></li></ul>
|
||||
<h4>✨ <a href="lora/index.html">Low-Rank Adaptation (LoRA)</a></h4>
|
||||
<h4>✨ <a href="neox/index.html">Eleuther GPT-NeoX</a></h4>
|
||||
<ul><li><a href="neox/samples/generate.html">Generate on a 48GB GPU</a> </li>
|
||||
<li><a href="neox/samples/finetune.html">Finetune on two 48GB GPUs</a> </li>
|
||||
|
||||
@ -283,14 +283,14 @@
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">83</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">12</span><span class="p">):</span>
|
||||
<span class="lineno">84</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_1.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.pre_norm.weight'</span>
|
||||
<span class="lineno">85</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_1.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.pre_norm.bias'</span>
|
||||
<span class="lineno">84</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_1.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn_norm.weight'</span>
|
||||
<span class="lineno">85</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_1.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn_norm.bias'</span>
|
||||
<span class="lineno">86</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.c_attn.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.qkv_projection.weight'</span>
|
||||
<span class="lineno">87</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.c_attn.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.qkv_projection.bias'</span>
|
||||
<span class="lineno">88</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.c_proj.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.output_projection.weight'</span>
|
||||
<span class="lineno">89</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.c_proj.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.attn.output_projection.bias'</span>
|
||||
<span class="lineno">90</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_2.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.post_norm.weight'</span>
|
||||
<span class="lineno">91</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_2.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.post_norm.bias'</span>
|
||||
<span class="lineno">90</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_2.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ffn_norm.weight'</span>
|
||||
<span class="lineno">91</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ln_2.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ffn_norm.bias'</span>
|
||||
<span class="lineno">92</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.mlp.c_fc.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ffn.linear_in.weight'</span>
|
||||
<span class="lineno">93</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.mlp.c_fc.bias'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ffn.linear_in.bias'</span>
|
||||
<span class="lineno">94</span> <span class="n">mapping</span><span class="p">[</span><span class="sa">f</span><span class="s1">'transformer.h.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.mlp.c_proj.weight'</span><span class="p">]</span> <span class="o">=</span> <span class="sa">f</span><span class="s1">'blocks.</span><span class="si">{</span><span class="n">i</span><span class="si">}</span><span class="s1">.ffn.linear_out.weight'</span>
|
||||
@ -340,40 +340,32 @@
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">113</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">load_state_dict</span><span class="p">(</span><span class="n">new_state_dict</span><span class="p">,</span> <span class="n">strict</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">113</span> <span class="n">missing_keys</span><span class="p">,</span> <span class="n">unexpected_keys</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">load_state_dict</span><span class="p">(</span><span class="n">new_state_dict</span><span class="p">,</span> <span class="n">strict</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-19'>
|
||||
<div class='docs doc-strings'>
|
||||
<div class='docs'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-19'>#</a>
|
||||
</div>
|
||||
<p>make sure that only lora weights are not loaded </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">116</span> <span class="k">assert</span> <span class="nb">all</span><span class="p">(</span><span class="s1">'lora'</span> <span class="ow">in</span> <span class="n">key</span> <span class="k">for</span> <span class="n">key</span> <span class="ow">in</span> <span class="n">missing_keys</span><span class="p">)</span>
|
||||
<span class="lineno">117</span> <span class="k">assert</span> <span class="ow">not</span> <span class="n">unexpected_keys</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-20'>
|
||||
<div class='docs doc-strings'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-20'>#</a>
|
||||
</div>
|
||||
<h3>Initialize the model, optimizer and dataloader</h3>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">115</span> <span class="k">def</span> <span class="nf">initialize</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-20'>
|
||||
<div class='docs'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-20'>#</a>
|
||||
</div>
|
||||
<p>Initialize the <a href="gpt2.html">GPT2 model</a> </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">120</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span> <span class="o">=</span> <span class="n">GPTModel</span><span class="p">(</span>
|
||||
<span class="lineno">121</span> <span class="n">layer_norm_epsilon</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">layer_norm_epsilon</span><span class="p">,</span>
|
||||
<span class="lineno">122</span> <span class="n">d_model</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">d_model</span><span class="p">,</span>
|
||||
<span class="lineno">123</span> <span class="n">n_layers</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">n_layers</span><span class="p">,</span>
|
||||
<span class="lineno">124</span> <span class="n">n_heads</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">n_heads</span><span class="p">,</span>
|
||||
<span class="lineno">125</span> <span class="n">n_positions</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">n_positions</span><span class="p">,</span>
|
||||
<span class="lineno">126</span> <span class="n">vocab_size</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">vocab_size</span><span class="p">,</span>
|
||||
<span class="lineno">127</span> <span class="n">r</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">lora_r</span><span class="p">,</span>
|
||||
<span class="lineno">128</span> <span class="p">)</span>
|
||||
<span class="lineno">129</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">119</span> <span class="k">def</span> <span class="nf">initialize</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-21'>
|
||||
@ -381,11 +373,20 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-21'>#</a>
|
||||
</div>
|
||||
<p>Load pre-trained model weights </p>
|
||||
<p>Initialize the <a href="gpt2.html">GPT2 model</a> </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">131</span> <span class="bp">self</span><span class="o">.</span><span class="n">_load_pretrained_weights</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">124</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span> <span class="o">=</span> <span class="n">GPTModel</span><span class="p">(</span>
|
||||
<span class="lineno">125</span> <span class="n">layer_norm_epsilon</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">layer_norm_epsilon</span><span class="p">,</span>
|
||||
<span class="lineno">126</span> <span class="n">d_model</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">d_model</span><span class="p">,</span>
|
||||
<span class="lineno">127</span> <span class="n">n_layers</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">n_layers</span><span class="p">,</span>
|
||||
<span class="lineno">128</span> <span class="n">n_heads</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">n_heads</span><span class="p">,</span>
|
||||
<span class="lineno">129</span> <span class="n">n_positions</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">n_positions</span><span class="p">,</span>
|
||||
<span class="lineno">130</span> <span class="n">vocab_size</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">vocab_size</span><span class="p">,</span>
|
||||
<span class="lineno">131</span> <span class="n">r</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">lora_r</span><span class="p">,</span>
|
||||
<span class="lineno">132</span> <span class="p">)</span>
|
||||
<span class="lineno">133</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-22'>
|
||||
@ -393,11 +394,11 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-22'>#</a>
|
||||
</div>
|
||||
<p>Initialize the optimizer </p>
|
||||
<p>Load pre-trained model weights </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">134</span> <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span> <span class="o">=</span> <span class="n">Adam</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">parameters</span><span class="p">(),</span> <span class="n">lr</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">learning_rate</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">135</span> <span class="bp">self</span><span class="o">.</span><span class="n">_load_pretrained_weights</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-23'>
|
||||
@ -405,34 +406,35 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-23'>#</a>
|
||||
</div>
|
||||
<p>Initialize the optimizer </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">138</span> <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span> <span class="o">=</span> <span class="n">Adam</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">parameters</span><span class="p">(),</span> <span class="n">lr</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">learning_rate</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-24'>
|
||||
<div class='docs'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-24'>#</a>
|
||||
</div>
|
||||
<p>Initialize the data loader </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">137</span> <span class="bp">self</span><span class="o">.</span><span class="n">data_loader</span> <span class="o">=</span> <span class="n">DataLoader</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">text</span><span class="p">,</span> <span class="n">batch_size</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">shuffle</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">141</span> <span class="bp">self</span><span class="o">.</span><span class="n">data_loader</span> <span class="o">=</span> <span class="n">DataLoader</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">text</span><span class="p">,</span> <span class="n">batch_size</span><span class="o">=</span><span class="bp">self</span><span class="o">.</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">shuffle</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-24'>
|
||||
<div class='section' id='section-25'>
|
||||
<div class='docs doc-strings'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-24'>#</a>
|
||||
<a href='#section-25'>#</a>
|
||||
</div>
|
||||
<h3>Training loop</h3>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">139</span> <span class="k">def</span> <span class="nf">run</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-25'>
|
||||
<div class='docs'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-25'>#</a>
|
||||
</div>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">144</span> <span class="k">for</span> <span class="n">_</span> <span class="ow">in</span> <span class="n">monit</span><span class="o">.</span><span class="n">loop</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">epochs</span><span class="p">):</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">143</span> <span class="k">def</span> <span class="nf">run</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-26'>
|
||||
@ -440,13 +442,10 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-26'>#</a>
|
||||
</div>
|
||||
<p><code class="highlight"><span></span><span class="n">inputs</span></code>
|
||||
has shape <code class="highlight"><span></span><span class="p">[</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">seq_len</span><span class="p">]</span></code>
|
||||
</p>
|
||||
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">146</span> <span class="k">for</span> <span class="p">(</span><span class="n">inputs</span><span class="p">,)</span> <span class="ow">in</span> <span class="n">monit</span><span class="o">.</span><span class="n">iterate</span><span class="p">(</span><span class="s1">'Train'</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">data_loader</span><span class="p">):</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">148</span> <span class="k">for</span> <span class="n">_</span> <span class="ow">in</span> <span class="n">monit</span><span class="o">.</span><span class="n">loop</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">epochs</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-27'>
|
||||
@ -454,12 +453,13 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-27'>#</a>
|
||||
</div>
|
||||
<p>Move <code class="highlight"><span></span><span class="n">inputs</span></code>
|
||||
to device </p>
|
||||
<p><code class="highlight"><span></span><span class="n">inputs</span></code>
|
||||
has shape <code class="highlight"><span></span><span class="p">[</span><span class="n">batch_size</span><span class="p">,</span> <span class="n">seq_len</span><span class="p">]</span></code>
|
||||
</p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">148</span> <span class="n">inputs</span> <span class="o">=</span> <span class="n">inputs</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">150</span> <span class="k">for</span> <span class="p">(</span><span class="n">inputs</span><span class="p">,)</span> <span class="ow">in</span> <span class="n">monit</span><span class="o">.</span><span class="n">iterate</span><span class="p">(</span><span class="s1">'Train'</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">data_loader</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-28'>
|
||||
@ -467,11 +467,12 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-28'>#</a>
|
||||
</div>
|
||||
<p>Call the model, with the all but the last token </p>
|
||||
<p>Move <code class="highlight"><span></span><span class="n">inputs</span></code>
|
||||
to device </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">150</span> <span class="n">logits</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="p">(</span><span class="n">inputs</span><span class="p">[:,</span> <span class="p">:</span><span class="o">-</span><span class="mi">1</span><span class="p">])</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">152</span> <span class="n">inputs</span> <span class="o">=</span> <span class="n">inputs</span><span class="o">.</span><span class="n">to</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">device</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-29'>
|
||||
@ -479,11 +480,11 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-29'>#</a>
|
||||
</div>
|
||||
<p>Get cross entropy loss </p>
|
||||
<p>Call the model, with the all but the last token </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">152</span> <span class="n">loss</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">loss_func</span><span class="p">(</span><span class="n">logits</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="n">logits</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]),</span> <span class="n">inputs</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">:]</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">))</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">154</span> <span class="n">logits</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">model</span><span class="p">(</span><span class="n">inputs</span><span class="p">[:,</span> <span class="p">:</span><span class="o">-</span><span class="mi">1</span><span class="p">])</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-30'>
|
||||
@ -491,11 +492,11 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-30'>#</a>
|
||||
</div>
|
||||
<p>Make gradients 0 </p>
|
||||
<p>Get cross entropy loss </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">155</span> <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="o">.</span><span class="n">zero_grad</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">156</span> <span class="n">loss</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">loss_func</span><span class="p">(</span><span class="n">logits</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="n">logits</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]),</span> <span class="n">inputs</span><span class="p">[:,</span> <span class="mi">1</span><span class="p">:]</span><span class="o">.</span><span class="n">reshape</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">))</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-31'>
|
||||
@ -503,11 +504,11 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-31'>#</a>
|
||||
</div>
|
||||
<p>Compute gradients </p>
|
||||
<p>Make gradients 0 </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">157</span> <span class="n">loss</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">159</span> <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="o">.</span><span class="n">zero_grad</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-32'>
|
||||
@ -515,11 +516,11 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-32'>#</a>
|
||||
</div>
|
||||
<p>Optimize </p>
|
||||
<p>Compute gradients </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">159</span> <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="o">.</span><span class="n">step</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">161</span> <span class="n">loss</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-33'>
|
||||
@ -527,12 +528,11 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-33'>#</a>
|
||||
</div>
|
||||
<p>Log the loss </p>
|
||||
<p>Optimize </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">162</span> <span class="n">tracker</span><span class="o">.</span><span class="n">save</span><span class="p">({</span><span class="s1">'loss'</span><span class="p">:</span> <span class="n">loss</span><span class="p">})</span>
|
||||
<span class="lineno">163</span> <span class="n">tracker</span><span class="o">.</span><span class="n">add_global_step</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">163</span> <span class="bp">self</span><span class="o">.</span><span class="n">optimizer</span><span class="o">.</span><span class="n">step</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-34'>
|
||||
@ -540,46 +540,59 @@
|
||||
<div class='section-link'>
|
||||
<a href='#section-34'>#</a>
|
||||
</div>
|
||||
<p>Log the loss </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">166</span> <span class="n">tracker</span><span class="o">.</span><span class="n">save</span><span class="p">({</span><span class="s1">'loss'</span><span class="p">:</span> <span class="n">loss</span><span class="p">})</span>
|
||||
<span class="lineno">167</span> <span class="n">tracker</span><span class="o">.</span><span class="n">add_global_step</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-35'>
|
||||
<div class='docs'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-35'>#</a>
|
||||
</div>
|
||||
<p> </p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">165</span> <span class="n">tracker</span><span class="o">.</span><span class="n">new_line</span><span class="p">()</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">169</span> <span class="n">tracker</span><span class="o">.</span><span class="n">new_line</span><span class="p">()</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-35'>
|
||||
<div class='section' id='section-36'>
|
||||
<div class='docs doc-strings'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-35'>#</a>
|
||||
<a href='#section-36'>#</a>
|
||||
</div>
|
||||
<h3>Tiny Shakespeare dataset</h3>
|
||||
<p>It will download from the url if not present</p>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">168</span><span class="nd">@option</span><span class="p">(</span><span class="n">Trainer</span><span class="o">.</span><span class="n">text</span><span class="p">)</span>
|
||||
<span class="lineno">169</span><span class="k">def</span> <span class="nf">tiny_shakespeare</span><span class="p">(</span><span class="n">c</span><span class="p">:</span> <span class="n">Trainer</span><span class="p">):</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">172</span><span class="nd">@option</span><span class="p">(</span><span class="n">Trainer</span><span class="o">.</span><span class="n">text</span><span class="p">)</span>
|
||||
<span class="lineno">173</span><span class="k">def</span> <span class="nf">tiny_shakespeare</span><span class="p">(</span><span class="n">c</span><span class="p">:</span> <span class="n">Trainer</span><span class="p">):</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='section' id='section-36'>
|
||||
<div class='section' id='section-37'>
|
||||
<div class='docs'>
|
||||
<div class='section-link'>
|
||||
<a href='#section-36'>#</a>
|
||||
<a href='#section-37'>#</a>
|
||||
</div>
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">175</span> <span class="n">path</span> <span class="o">=</span> <span class="n">lab</span><span class="o">.</span><span class="n">get_data_path</span><span class="p">()</span> <span class="o">/</span> <span class="s1">'tiny_shakespeare.txt'</span>
|
||||
<span class="lineno">176</span> <span class="k">if</span> <span class="ow">not</span> <span class="n">path</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
|
||||
<span class="lineno">177</span> <span class="n">download_file</span><span class="p">(</span><span class="s2">"https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt"</span><span class="p">,</span> <span class="n">path</span><span class="p">)</span>
|
||||
<span class="lineno">178</span> <span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="s1">'r'</span><span class="p">,</span> <span class="n">encoding</span><span class="o">=</span><span class="s1">'utf-8'</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
|
||||
<span class="lineno">179</span> <span class="n">text</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">read</span><span class="p">()</span>
|
||||
<span class="lineno">180</span>
|
||||
<span class="lineno">181</span> <span class="n">tokens</span> <span class="o">=</span> <span class="n">c</span><span class="o">.</span><span class="n">tokenizer</span><span class="o">.</span><span class="n">encode</span><span class="p">(</span><span class="n">text</span><span class="p">)</span>
|
||||
<span class="lineno">182</span> <span class="n">num_batches</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">tokens</span><span class="p">)</span> <span class="o">//</span> <span class="p">(</span><span class="n">c</span><span class="o">.</span><span class="n">batch_size</span> <span class="o">*</span> <span class="n">c</span><span class="o">.</span><span class="n">context_len</span><span class="p">)</span>
|
||||
<span class="lineno">183</span> <span class="n">tokens</span> <span class="o">=</span> <span class="n">tokens</span><span class="p">[:</span><span class="n">num_batches</span> <span class="o">*</span> <span class="n">c</span><span class="o">.</span><span class="n">batch_size</span> <span class="o">*</span> <span class="n">c</span><span class="o">.</span><span class="n">context_len</span><span class="p">]</span>
|
||||
<span class="lineno">184</span> <span class="n">input_ids</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">(</span><span class="n">tokens</span><span class="p">)</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="n">c</span><span class="o">.</span><span class="n">context_len</span><span class="p">)</span>
|
||||
<span class="lineno">185</span> <span class="k">return</span> <span class="n">TensorDataset</span><span class="p">(</span><span class="n">input_ids</span><span class="p">)</span></pre></div>
|
||||
<div class="highlight"><pre><span class="lineno">179</span> <span class="n">path</span> <span class="o">=</span> <span class="n">lab</span><span class="o">.</span><span class="n">get_data_path</span><span class="p">()</span> <span class="o">/</span> <span class="s1">'tiny_shakespeare.txt'</span>
|
||||
<span class="lineno">180</span> <span class="k">if</span> <span class="ow">not</span> <span class="n">path</span><span class="o">.</span><span class="n">exists</span><span class="p">():</span>
|
||||
<span class="lineno">181</span> <span class="n">download_file</span><span class="p">(</span><span class="s2">"https://raw.githubusercontent.com/karpathy/char-rnn/master/data/tinyshakespeare/input.txt"</span><span class="p">,</span> <span class="n">path</span><span class="p">)</span>
|
||||
<span class="lineno">182</span> <span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="n">path</span><span class="p">,</span> <span class="s1">'r'</span><span class="p">,</span> <span class="n">encoding</span><span class="o">=</span><span class="s1">'utf-8'</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
|
||||
<span class="lineno">183</span> <span class="n">text</span> <span class="o">=</span> <span class="n">f</span><span class="o">.</span><span class="n">read</span><span class="p">()</span>
|
||||
<span class="lineno">184</span>
|
||||
<span class="lineno">185</span> <span class="n">tokens</span> <span class="o">=</span> <span class="n">c</span><span class="o">.</span><span class="n">tokenizer</span><span class="o">.</span><span class="n">encode</span><span class="p">(</span><span class="n">text</span><span class="p">)</span>
|
||||
<span class="lineno">186</span> <span class="n">num_batches</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">tokens</span><span class="p">)</span> <span class="o">//</span> <span class="p">(</span><span class="n">c</span><span class="o">.</span><span class="n">batch_size</span> <span class="o">*</span> <span class="n">c</span><span class="o">.</span><span class="n">context_len</span><span class="p">)</span>
|
||||
<span class="lineno">187</span> <span class="n">tokens</span> <span class="o">=</span> <span class="n">tokens</span><span class="p">[:</span><span class="n">num_batches</span> <span class="o">*</span> <span class="n">c</span><span class="o">.</span><span class="n">batch_size</span> <span class="o">*</span> <span class="n">c</span><span class="o">.</span><span class="n">context_len</span><span class="p">]</span>
|
||||
<span class="lineno">188</span> <span class="n">input_ids</span> <span class="o">=</span> <span class="n">torch</span><span class="o">.</span><span class="n">tensor</span><span class="p">(</span><span class="n">tokens</span><span class="p">)</span><span class="o">.</span><span class="n">view</span><span class="p">(</span><span class="o">-</span><span class="mi">1</span><span class="p">,</span> <span class="n">c</span><span class="o">.</span><span class="n">context_len</span><span class="p">)</span>
|
||||
<span class="lineno">189</span> <span class="k">return</span> <span class="n">TensorDataset</span><span class="p">(</span><span class="n">input_ids</span><span class="p">)</span></pre></div>
|
||||
</div>
|
||||
</div>
|
||||
<div class='footer'>
|
||||
|
||||
@ -230,7 +230,7 @@
|
||||
|
||||
</div>
|
||||
<div class='code'>
|
||||
<div class="highlight"><pre><span class="lineno">82</span> <span class="bp">self</span><span class="o">.</span><span class="n">lora_b</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Parameter</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">empty</span><span class="p">((</span><span class="n">outfeatures</span><span class="p">,</span> <span class="n">r</span><span class="p">)))</span>
|
||||
<div class="highlight"><pre><span class="lineno">82</span> <span class="bp">self</span><span class="o">.</span><span class="n">lora_b</span> <span class="o">=</span> <span class="n">nn</span><span class="o">.</span><span class="n">Parameter</span><span class="p">(</span><span class="n">torch</span><span class="o">.</span><span class="n">empty</span><span class="p">((</span><span class="n">out_features</span><span class="p">,</span> <span class="n">r</span><span class="p">)))</span>
|
||||
<span class="lineno">83</span>
|
||||
<span class="lineno">84</span> <span class="k">with</span> <span class="n">torch</span><span class="o">.</span><span class="n">no_grad</span><span class="p">():</span></pre></div>
|
||||
</div>
|
||||
|
||||
@ -1303,14 +1303,14 @@
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/lora/index.html</loc>
|
||||
<lastmod>2024-08-03T16:30:00+00:00</lastmod>
|
||||
<lastmod>2024-08-23T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
<url>
|
||||
<loc>https://nn.labml.ai/lora/experiment.html</loc>
|
||||
<lastmod>2024-08-18T16:30:00+00:00</lastmod>
|
||||
<lastmod>2024-08-23T16:30:00+00:00</lastmod>
|
||||
<priority>1.00</priority>
|
||||
</url>
|
||||
|
||||
|
||||
Reference in New Issue
Block a user