rollback user guide

docs/en/advance/long_context.md

@@ -96,9 +96,11 @@ This test takes approximately 364 seconds per round when conducted on A100-80G G
 The following codes demonstrate how to use LMDeploy to calculate perplexity.
 
 ```python
+from transformers import AutoTokenizer
 from lmdeploy import TurbomindEngineConfig, pipeline
+import numpy as np
 
-# build pipeline
+# load model and tokenizer
 model_repoid_or_path = 'internlm/internlm2_5-7b-chat-1m'
 backend_config = TurbomindEngineConfig(
         rope_scaling_factor=2.5,
@@ -107,9 +109,11 @@ backend_config = TurbomindEngineConfig(
         cache_max_entry_count=0.7,
         tp=4)
 pipe = pipeline(model_repoid_or_path, backend_config=backend_config)
+tokenizer = AutoTokenizer.from_pretrained(model_repoid_or_path, trust_remote_code=True)
 
 # get perplexity
 text = 'Use a long prompt to replace this sentence'
-ppl = pipe.get_ppl(text)
+input_ids = tokenizer.encode(text)
+ppl = pipe.get_ppl(input_ids)[0]
 print(ppl)
 ```

docs/en/llm/pipeline.md

@@ -119,21 +119,21 @@ for item in pipe.stream_infer(prompts, gen_config=gen_config):
 - **An example to cauculate logits & ppl:**
 
 ```python
+from transformers import AutoTokenizer
 from lmdeploy import pipeline
-
 model_repoid_or_path='internlm/internlm2_5-7b-chat'
 pipe = pipeline(model_repoid_or_path)
-
-prompts = [
-    "Hello, I am an AI assistant named InternLM. I am developed by Shanghai AI Laboratory",
-    "How to use LMDeploy to deploy a LLM model?"
-]
+tokenizer = AutoTokenizer.from_pretrained(model_repoid_or_path, trust_remote_code=True)
 
 # logits
-logits = pipe.get_logits(prompts)
+messages = [
+   {"role": "user", "content": "Hello, how are you?"},
+]
+input_ids = tokenizer.apply_chat_template(messages)
+logits = pipe.get_logits(input_ids)
 
 # ppl
-ppl = pipe.get_ppl(prompts)
+ppl = pipe.get_ppl(input_ids)
 ```
 
 - **Below is an example for pytorch backend. Please install triton first.**

docs/zh_cn/advance/long_context.md

@@ -96,9 +96,11 @@ passkey_retrieval(session_len, 5)
 下面展示使用 LMDeploy 计算困惑度的用法
 
 ```python
+from transformers import AutoTokenizer
 from lmdeploy import TurbomindEngineConfig, pipeline
+import numpy as np
 
-# build pipeline
+# load model and tokenizer
 model_repoid_or_path = 'internlm/internlm2_5-7b-chat-1m'
 backend_config = TurbomindEngineConfig(
         rope_scaling_factor=2.5,
@@ -107,9 +109,11 @@ backend_config = TurbomindEngineConfig(
         cache_max_entry_count=0.7,
         tp=4)
 pipe = pipeline(model_repoid_or_path, backend_config=backend_config)
+tokenizer = AutoTokenizer.from_pretrained(model_repoid_or_path, trust_remote_code=True)
 
 # get perplexity
 text = 'Use a long prompt to replace this sentence'
-loss = pipe.get_ppl(text)
+input_ids = tokenizer.encode(text)
+loss = pipe.get_ppl(input_ids)[0]
 print(ppl)
 ```

docs/zh_cn/llm/pipeline.md

@@ -119,21 +119,21 @@ for item in pipe.stream_infer(prompts, gen_config=gen_config):
 - **计算 logits & ppl:**
 
 ```python
+from transformers import AutoTokenizer
 from lmdeploy import pipeline
-
 model_repoid_or_path='internlm/internlm2_5-7b-chat'
 pipe = pipeline(model_repoid_or_path)
-
-prompts = [
-    "Hello, I am an AI assistant named InternLM. I am developed by Shanghai AI Laboratory",
-    "How to use LMDeploy to deploy a LLM model?"
-]
+tokenizer = AutoTokenizer.from_pretrained(model_repoid_or_path, trust_remote_code=True)
 
 # logits
-logits = pipe.get_logits(prompts)
+messages = [
+   {"role": "user", "content": "Hello, how are you?"},
+]
+input_ids = tokenizer.apply_chat_template(messages)
+logits = pipe.get_logits(input_ids)
 
 # ppl
-ppl = pipe.get_ppl(prompts)
+ppl = pipe.get_ppl(input_ids)
 ```
 
 - **使用 pytorch 后端**