Buckets:

hf-doc-build
/

doc-dev

Files

xet

hf-doc-build/doc-dev / openenv /pr_749 /en /tutorials /browsergym-harness.html

HuggingFaceDocBuilder

30 days ago

download

raw

26 kB

	<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"BrowserGym Harness Rollouts","local":"browsergym-harness-rollouts","sections":[{"title":"What You’ll Build","local":"what-youll-build","sections":[],"depth":2},{"title":"Install Dependencies","local":"install-dependencies","sections":[],"depth":2},{"title":"Build The Session Factory","local":"build-the-session-factory","sections":[],"depth":2},{"title":"Wrap TRL Generation","local":"wrap-trl-generation","sections":[],"depth":2},{"title":"Create The Rollout Function","local":"create-the-rollout-function","sections":[],"depth":2},{"title":"Full Example","local":"full-example","sections":[],"depth":2}],"depth":1}">
	<link href="/docs/openenv/pr_749/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/entry/start.85477f45.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/scheduler.2b22cead.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/singletons.63566282.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/paths.dd876c7b.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/entry/app.51835dc5.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/preload-helper.0820fbc7.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/index.1a0e8013.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/nodes/0.167255c0.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/each.e59479a4.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/nodes/57.4355ff99.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/Heading.c0d3f116.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.21bcf336.js">
	<link rel="modulepreload" href="/docs/openenv/pr_749/en/_app/immutable/chunks/CodeBlock.c8d73295.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"BrowserGym Harness Rollouts","local":"browsergym-harness-rollouts","sections":[{"title":"What You’ll Build","local":"what-youll-build","sections":[],"depth":2},{"title":"Install Dependencies","local":"install-dependencies","sections":[],"depth":2},{"title":"Build The Session Factory","local":"build-the-session-factory","sections":[],"depth":2},{"title":"Wrap TRL Generation","local":"wrap-trl-generation","sections":[],"depth":2},{"title":"Create The Rollout Function","local":"create-the-rollout-function","sections":[],"depth":2},{"title":"Full Example","local":"full-example","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"><div class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg></span> <span>Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg></button></div> </div> <h1 class="relative group"><a id="browsergym-harness-rollouts" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#browsergym-harness-rollouts"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>BrowserGym Harness Rollouts</span></h1> <p data-svelte-h="svelte-1540o3q">This tutorial shows how to drive BrowserGym through the OpenEnv harness runtime
	when a trainer needs to keep token sampling, logprobs, and reward assignment
	inside the training loop.</p> <blockquote class="note" data-svelte-h="svelte-57zv20"><p>Use this pattern for tool-driven BrowserGym rollouts. For a standard
	<code>reset()</code> / <code>step()</code> GRPO flow, keep using the Wordle GRPO tutorial.</p></blockquote> <h2 class="relative group"><a id="what-youll-build" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#what-youll-build"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>What You’ll Build</span></h2> <ul data-svelte-h="svelte-62gmut"><li>A BrowserGym session factory that creates one environment client per rollout.</li> <li>A harness rollout function that TRL can call during training.</li> <li>A model-step wrapper that converts generated BrowserGym action text into
	structured tool calls.</li></ul> <h2 class="relative group"><a id="install-dependencies" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#install-dependencies"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Install Dependencies</span></h2> <p data-svelte-h="svelte-1ynlc70">Install OpenEnv, TRL, and the BrowserGym environment package:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class="language-bash "><!-- HTML_TAG_START -->pip install -U <span class="hljs-string">"trl[vllm]"</span> peft trackio kernels
	pip install -U git+https://github.com/huggingface/OpenEnv.git
	pip install -U <span class="hljs-string">"openenv-browsergym @ git+https://huggingface.co/spaces/openenv/browsergym_env"</span><!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="build-the-session-factory" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#build-the-session-factory"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Build The Session Factory</span></h2> <p data-svelte-h="svelte-1qg1a9f"><code>BrowserGymSessionFactory</code> adapts a BrowserGym client into the harness
	<code>ResourceSession</code> interface. If your training setup already has an
	<code>environment_factory</code>, pass that factory as <code>client_factory</code> so every rollout
	gets a fresh environment instance.</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class="language-python "><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> browsergym_env <span class="hljs-keyword">import</span> BrowserGymEnv
	<span class="hljs-keyword">from</span> browsergym_env.harness <span class="hljs-keyword">import</span> BrowserGymSessionFactory

	space_url = <span class="hljs-string">"https://openenv-browsergym-env.hf.space"</span>


	<span class="hljs-keyword">def</span> <span class="hljs-title function_">environment_factory</span>():
	<span class="hljs-keyword">return</span> BrowserGymEnv(base_url=space_url)


	session_factory = BrowserGymSessionFactory(
	client_factory=environment_factory,
	default_task=<span class="hljs-string">"click-test"</span>,
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1b3uzy8">The session exposes BrowserGym actions such as <code>click</code>, <code>fill</code>, <code>send_keys</code>,
	<code>scroll</code>, and <code>noop</code> as MCP-style tools while still executing the corresponding
	BrowserGym action strings under the hood.</p> <h2 class="relative group"><a id="wrap-trl-generation" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#wrap-trl-generation"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Wrap TRL Generation</span></h2> <p data-svelte-h="svelte-lnvo5a">The harness calls a <code>model_step</code> function for each turn. The model step should
	use the trainer-owned generation path, then return a <code>ModelStepResult</code> with the
	completion text, token ids, logprobs, and exactly one BrowserGym tool call.</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class="language-python "><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> browsergym_env.harness <span class="hljs-keyword">import</span> build_browsergym_action_tool_call
	<span class="hljs-keyword">from</span> openenv.core.harness <span class="hljs-keyword">import</span> ModelStepResult
	<span class="hljs-keyword">from</span> openenv.core.llm_client <span class="hljs-keyword">import</span> LLMResponse
	<span class="hljs-keyword">from</span> trl.experimental.openenv <span class="hljs-keyword">import</span> generate_rollout_completions


	<span class="hljs-keyword">def</span> <span class="hljs-title function_">build_trl_browsergym_model_step</span>(<span class="hljs-params">trainer, tokenizer</span>):
	<span class="hljs-keyword">def</span> <span class="hljs-title function_">model_step</span>(<span class="hljs-params">messages, tools, sampling</span>):
	<span class="hljs-keyword">del</span> tools, sampling
	prompt_text = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=<span class="hljs-literal">True</span>,
	tokenize=<span class="hljs-literal">False</span>,
	)
	rollout_output = generate_rollout_completions(trainer, [prompt_text])[<span class="hljs-number">0</span>]
	completion_text = rollout_output.get(<span class="hljs-string">"text"</span>) <span class="hljs-keyword">or</span> tokenizer.decode(
	rollout_output[<span class="hljs-string">"completion_ids"</span>],
	skip_special_tokens=<span class="hljs-literal">True</span>,
	)
	tool_call = build_browsergym_action_tool_call(completion_text)
	<span class="hljs-keyword">return</span> ModelStepResult(
	response=LLMResponse(content=completion_text, tool_calls=[tool_call]),
	prompt_ids=<span class="hljs-built_in">list</span>(rollout_output[<span class="hljs-string">"prompt_ids"</span>]),
	completion_ids=<span class="hljs-built_in">list</span>(rollout_output[<span class="hljs-string">"completion_ids"</span>]),
	logprobs=<span class="hljs-built_in">list</span>(rollout_output[<span class="hljs-string">"logprobs"</span>]),
	)

	<span class="hljs-keyword">return</span> model_step<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1l5v3lb">In practice, you should add a small parser around the completion text so common
	outputs like <code>Action: click('13')</code> are normalized before calling
	<code>build_browsergym_action_tool_call</code>.</p> <h2 class="relative group"><a id="create-the-rollout-function" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#create-the-rollout-function"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Create The Rollout Function</span></h2> <p data-svelte-h="svelte-vbg4ew">Pass the session factory, white-box harness adapter, and model-step builder to
	<code>build_harness_rollout_func</code>:</p> <div class="code-block relative "><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class="language-python "><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> openenv.core.harness <span class="hljs-keyword">import</span> (
	HarnessRunLimits,
	MCPHarnessAdapter,
	build_harness_rollout_func,
	)

	rollout_func = build_harness_rollout_func(
	session_factory=session_factory,
	harness_adapter=MCPHarnessAdapter(),
	model_step_builder=<span class="hljs-keyword">lambda</span> trainer, session: build_trl_browsergym_model_step(
	trainer,
	tokenizer,
	),
	limits=HarnessRunLimits(max_turns=<span class="hljs-number">10</span>),
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1337m4t">The returned function accepts TRL prompts and a trainer, runs one harness-backed
	BrowserGym episode per prompt, and returns <code>prompt_ids</code>, <code>completion_ids</code>,
	<code>logprobs</code>, <code>env_reward</code>, and <code>verify_metrics</code>.</p> <h2 class="relative group"><a id="full-example" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#full-example"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Full Example</span></h2> <p data-svelte-h="svelte-bh4c0y">See <a href="https://github.com/huggingface/OpenEnv/blob/main/examples/browsergym_harness.py" rel="nofollow"><code>examples/browsergym_harness.py</code></a>
	for a complete TRL-oriented helper that includes action normalization and a
	ready-to-use <code>build_browsergym_rollout_func</code>.</p> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/openenv/blob/main/docs/source/tutorials/browsergym-harness.md" target="_blank"><svg class="mr-1" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M31,16l-7,7l-1.41-1.41L28.17,16l-5.58-5.59L24,9l7,7z"></path><path d="M1,16l7-7l1.41,1.41L3.83,16l5.58,5.59L8,23l-7-7z"></path><path d="M12.419,25.484L17.639,6.552l1.932,0.518L14.351,26.002z"></path></svg> <span data-svelte-h="svelte-zjs2n5"><span class="underline">Update</span> on GitHub</span></a> <p></p>

	<script>
	{
	__sveltekit_1qwoa43 = {
	assets: "/docs/openenv/pr_749/en",
	base: "/docs/openenv/pr_749/en",
	env: {}
	};

	const element = document.currentScript.parentElement;

	const data = [null,null];

	Promise.all([
	import("/docs/openenv/pr_749/en/_app/immutable/entry/start.85477f45.js"),
	import("/docs/openenv/pr_749/en/_app/immutable/entry/app.51835dc5.js")
	]).then(([kit, app]) => {
	kit.start(app, element, {
	node_ids: [0, 57],
	data,
	form: null,
	error: null
	});
	});
	}
	</script>

Xet Storage Details

Size:: 26 kB
Xet hash:: 469786ee37d036d10eda24c5504f3ebde3669dc9f16a2ab0095d1f8d9908dfaa

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.