diff --git a/README.md b/README.md
index 0133fa5..8ba93f6 100644
--- a/README.md
+++ b/README.md
@@ -192,6 +192,11 @@ by h g wells i the time traveller for so it will be convenient to speak of him w
 -----------------
 ```
 
+### pre-trained lm model
+
+This model was trained for 7 epochs using the full text of The Time Machine novel.
+[model](https://cpp-transformer-1252366230.cos.ap-beijing.myqcloud.com/lm/checkpoint_20250609_123600_7.bin)
+
 ## legacy version
 
 [v1](https://github.com/freelw/cpp-transformer/tree/v1_freeze_20250529)
diff --git a/dataloaders/language_model/lm_dataloader.cpp b/dataloaders/language_model/lm_dataloader.cpp
index 5143748..44cd6ef 100644
--- a/dataloaders/language_model/lm_dataloader.cpp
+++ b/dataloaders/language_model/lm_dataloader.cpp
@@ -32,6 +32,7 @@ void LMDataLoader::get_token_ids(
     }
 
     int token_ids_size = std::min((int)token_ids.size(), max_token_ids_size);
+    std::cout << "token_ids_size : " << token_ids_size << std::endl;
 
     for (size_t i = 0; i < token_ids_size; ++i) {
         std::vector<uint> src_step_tokens;
diff --git a/lm.cpp b/lm.cpp
index 4603179..e057163 100644
--- a/lm.cpp
+++ b/lm.cpp
@@ -102,10 +102,11 @@ int main(int argc, char* argv[]) {
     int gpu = 1;
     int max_words_cnt = 256;
     float lr = 0.001f;
+    int lm_predict_cnt = LM_PREDICT_CNT;
     std::string checkpoint;
     std::string corpus = TIMEMACHINE_RESOURCE_NAME;
 
-    while ((opt = getopt(argc, argv, "f:c:e:l:b:g:m:")) != -1) {
+    while ((opt = getopt(argc, argv, "f:c:e:l:b:g:m:p:")) != -1) {
         switch (opt) {
         case 'f':
             corpus = optarg;
@@ -128,6 +129,9 @@ int main(int argc, char* argv[]) {
         case 'm':
             max_words_cnt = atoi(optarg);
             break;
+        case 'p':
+            lm_predict_cnt = atoi(optarg);
+            break;
         default:
             std::cerr << "Usage: " << argv[0]
                 << " -f <corpus> -c <checpoint> -e <epochs>" << std::endl;
@@ -249,7 +253,7 @@ int main(int argc, char* argv[]) {
             float* res_buffer = static_cast<float*>(::malloc(
                 res->get_tensor()->size()
             ));
-            for (int i = 0; i < LM_PREDICT_CNT; ++i) {
+            for (int i = 0; i < lm_predict_cnt; ++i) {
                 for (int j = 0; j < num_steps; ++j) {
                     tgt_token_ids_buffer[j] = src_token_ids[j];
                 }
diff --git a/test_lm.txt b/test_lm.txt
index 3592ed8..a1e6558 100644
--- a/test_lm.txt
+++ b/test_lm.txt
@@ -1 +1,2 @@
-the time machine
\ No newline at end of file
+the time machine
+the fire burned brightly