fix internlm-20b-chat model_id (#88)

Author: Jintao-Huang

README.md

@@ -44,7 +44,7 @@ Users can check the [documentation of Swift](docs/source/GetStarted/Introduction
    4. chatglm2 series: [chatglm2-6b](https://modelscope.cn/models/ZhipuAI/chatglm2-6b/summary), chatglm2-6b-32k
    5. llama series: llama2-7b, llama2-7b-chat, llama2-13b, llama2-13b-chat, llama2-70b, [llama2-70b-chat](https://modelscope.cn/models/modelscope/Llama-2-70b-chat-ms/summary)
    6. openbuddy-llama series: openbuddy-llama2-13b, openbuddy-llama-65b, [openbuddy-llama2-70b](https://modelscope.cn/models/OpenBuddy/openbuddy-llama2-70b-v10.1-bf16/summary)
-   7. internlm series: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b-chat/summary)
+   7. internlm series: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-chat-20b/summary)
    8. other: [polylm-13b](https://modelscope.cn/models/damo/nlp_polylm_13b_text_generation/summary), [seqgpt-560m](https://modelscope.cn/models/damo/nlp_seqgpt-560m/summary)
 3. supported features: quantization, DDP, model parallelism(device map), gradient checkpointing, gradient accumulation, pushing to modelscope hub, custom datasets, multimodal and agent SFT, mutli-round chat, ...
 4. supported datasets:

README_CN.md

@@ -42,7 +42,7 @@ SWIFT（Scalable lightWeight Infrastructure for Fine-Tuning）是一个可扩展
    4. chatglm2 系列: [chatglm2-6b](https://modelscope.cn/models/ZhipuAI/chatglm2-6b/summary), chatglm2-6b-32k
    5. llama 系列: llama2-7b, llama2-7b-chat, llama2-13b, llama2-13b-chat, llama2-70b, [llama2-70b-chat](https://modelscope.cn/models/modelscope/Llama-2-70b-chat-ms/summary)
    6. openbuddy-llama 系列: openbuddy-llama2-13b, openbuddy-llama-65b, [openbuddy-llama2-70b](https://modelscope.cn/models/OpenBuddy/openbuddy-llama2-70b-v10.1-bf16/summary)
-   7. internlm 系列: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b-chat/summary)
+   7. internlm 系列: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-chat-20b/summary)
    8. other: [polylm-13b](https://modelscope.cn/models/damo/nlp_polylm_13b_text_generation/summary), [seqgpt-560m](https://modelscope.cn/models/damo/nlp_seqgpt-560m/summary)
 3. 支持的特性: 模型量化, DDP, 模型并行(device_map), gradient checkpointing, 梯度累加, 支持推送ModelScope Hub, 自定义数据集, 多模态和Agent SFT, 多轮对话, ...
 4. 支持的数据集:

examples/pytorch/llm/README.md

@@ -23,7 +23,7 @@
    4. chatglm2 series: [chatglm2-6b](https://modelscope.cn/models/ZhipuAI/chatglm2-6b/summary), chatglm2-6b-32k
    5. llama series: llama2-7b, llama2-7b-chat, llama2-13b, llama2-13b-chat, llama2-70b, [llama2-70b-chat](https://modelscope.cn/models/modelscope/Llama-2-70b-chat-ms/summary)
    6. openbuddy-llama series: openbuddy-llama2-13b, openbuddy-llama-65b, [openbuddy-llama2-70b](https://modelscope.cn/models/OpenBuddy/openbuddy-llama2-70b-v10.1-bf16/summary)
-   7. internlm series: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b-chat/summary)
+   7. internlm series: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-chat-20b/summary)
    8. other: [polylm-13b](https://modelscope.cn/models/damo/nlp_polylm_13b_text_generation/summary), [seqgpt-560m](https://modelscope.cn/models/damo/nlp_seqgpt-560m/summary)
 3. supported features: quantization, DDP, model parallelism(device map), gradient checkpointing, gradient accumulation, pushing to modelscope hub, custom datasets, multimodal and agent SFT, mutli-round chat, ...
 4. supported datasets:
@@ -87,7 +87,7 @@ bash scripts/qwen_7b_chat/lora_ddp/infer.sh
 bash scripts/qwen_7b_chat/lora_mp_ddp/sft.sh
 bash scripts/qwen_7b_chat/lora_mp_ddp/infer.sh
 
-# sft(qlora) and infer qwen-7b-chat, Requires 12GB GPU memory.
+# sft(qlora) and infer qwen-7b-chat, Requires 10GB GPU memory.
 # If you want to use quantification, you need to `pip install bitsandbytes -U`
 # Recommended experimental environment: V100, A10, 3090
 bash scripts/qwen_7b_chat/qlora/sft.sh

examples/pytorch/llm/README_CN.md

@@ -24,7 +24,7 @@
    4. chatglm2 系列: [chatglm2-6b](https://modelscope.cn/models/ZhipuAI/chatglm2-6b/summary), chatglm2-6b-32k
    5. llama 系列: llama2-7b, llama2-7b-chat, llama2-13b, llama2-13b-chat, llama2-70b, [llama2-70b-chat](https://modelscope.cn/models/modelscope/Llama-2-70b-chat-ms/summary)
    6. openbuddy-llama 系列: openbuddy-llama2-13b, openbuddy-llama-65b, [openbuddy-llama2-70b](https://modelscope.cn/models/OpenBuddy/openbuddy-llama2-70b-v10.1-bf16/summary)
-   7. internlm 系列: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b-chat/summary)
+   7. internlm 系列: internlm-7b, internlm-7b-chat, internlm-7b-chat-8k, [internlm-20b](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-20b/summary), [internlm-20b-chat](https://modelscope.cn/models/Shanghai_AI_Laboratory/internlm-chat-20b/summary)
    8. other: [polylm-13b](https://modelscope.cn/models/damo/nlp_polylm_13b_text_generation/summary), [seqgpt-560m](https://modelscope.cn/models/damo/nlp_seqgpt-560m/summary)
 3. 支持的特性: 模型量化, DDP, 模型并行(device_map), gradient checkpointing, 梯度累加, 支持推送ModelScope Hub, 自定义数据集, 多模态和Agent SFT, 多轮对话, ...
 4. 支持的数据集:
@@ -89,7 +89,7 @@ bash scripts/qwen_7b_chat/lora_ddp/infer.sh
 bash scripts/qwen_7b_chat/lora_mp_ddp/sft.sh
 bash scripts/qwen_7b_chat/lora_mp_ddp/infer.sh
 
-# 微调(qlora)+推理 qwen-7b-chat, 需要12GB显存.
+# 微调(qlora)+推理 qwen-7b-chat, 需要10GB显存.
 # 如果你想要使用量化, 你需要`pip install bitsandbytes -U`
 # 推荐的实验环境: V100, 3090, A10
 bash scripts/qwen_7b_chat/qlora/sft.sh

examples/pytorch/llm/scripts/qwen_7b_chat/qlora_ddp/sft.sh

@@ -22,7 +22,7 @@ torchrun \
     --lora_alpha 32 \
     --lora_dropout_p 0. \
     --lora_target_modules ALL \
-    --gradient_checkpointing false \
+    --gradient_checkpointing true \
     --batch_size 1 \
     --weight_decay 0. \
     --learning_rate 1e-4 \

examples/pytorch/llm/src/utils/model.py

@@ -429,7 +429,7 @@ class ResTunerTM(NamedTuple):
         'lora_TM': LoRATM.internlm,
     },
     'internlm-20b-chat': {
-        'model_id': 'Shanghai_AI_Laboratory/internlm-20b-chat',
+        'model_id': 'Shanghai_AI_Laboratory/internlm-chat-20b',
         'revision': 'v1.0.0',
         'template': 'internlm',
         'lora_TM': LoRATM.internlm,