Upload ChatGLMForConditionalGeneration

README.md

@@ -1,7 +1,7 @@
 ---
-library_name: transformers
 language:
 - ko
+library_name: transformers
 ---
 
 # Model Card for Model ID

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "/workspace/new_korean_llama/src/result/4n3mone/glm-4-ko-9b-chat",
+  "_name_or_path": "/workspace/new_korean_llama/src/checkpoints/checkpoint-600",
   "add_bias_linear": false,
   "add_qkv_bias": true,
   "apply_query_key_layer_scaling": true,
@@ -44,7 +44,7 @@
   "seq_length": 131072,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.40.1",
+  "transformers_version": "4.41.2",
   "use_cache": false,
   "vocab_size": 151552
 }

generation_config.json

@@ -9,5 +9,5 @@
   "pad_token_id": 151329,
   "temperature": 0.8,
   "top_p": 0.8,
-  "transformers_version": "4.40.1"
+  "transformers_version": "4.41.2"
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a887940e5c42a8589fe8554dac25e23ebedf9ff44836ac52570a4aa0b197dcf
+oid sha256:64e7837813fb4760a1a02980a5809cbf27fef423994956f67f52355f457f96c8
 size 4984147224

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2a20f6c8376be09812240d2617e1c6555750d30c5799f2c841c0d9ab5beb9a41
+oid sha256:25b9bd8a1dde48779d341f11e451318090c9c8e8850407e0a16b7cc60b43a123
 size 4895071360

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:88d08031e5a7c351d207a09a9ed557fb28df038fbc03ea5f2d4034e89c256f61
+oid sha256:361736b8608890e9af90b90a3a3ec0882259daf5267c9194a771e4735832e9e5
 size 4895071384

model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e211c47324ba455c4680b973b7f04bd60e9c4c0005e1bf5805c0d6aefa7dd3e9
+oid sha256:87f3f87e9c2f70809a691a8768dfab04a4540549e8ec3cc9144e42e08d416dc2
 size 4025651256

modeling_chatglm.py

@@ -21,16 +21,20 @@ from transformers.modeling_outputs import (
     SequenceClassifierOutputWithPast,
 )
 from transformers.modeling_utils import PreTrainedModel
-from transformers.utils import logging, is_torch_npu_available, is_flash_attn_greater_or_equal_2_10, \
-    is_flash_attn_2_available
+from transformers.utils import logging, is_torch_npu_available
 from transformers.generation.logits_process import LogitsProcessor
 from transformers.generation.utils import LogitsProcessorList, StoppingCriteriaList, GenerationConfig, ModelOutput
 
 from .configuration_chatglm import ChatGLMConfig
 
-if is_flash_attn_2_available():
-    from flash_attn import flash_attn_func, flash_attn_varlen_func
-    from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa
+try:
+    from transformers.utils import is_flash_attn_greater_or_equal_2_10, is_flash_attn_2_available
+    if is_flash_attn_2_available():
+        from flash_attn import flash_attn_func, flash_attn_varlen_func
+        from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa
+except:
+    pass
+
 
 # flags required to enable jit fusion kernels